A fully autonomous high-throughput smart transmission electron microscope analysis system and method

Abstract

The application relates to a kind of full autonomous high-flux intelligent transmission electron microscope analysis system and method, belong to transmission electron microscope intelligent analysis technical field.It includes task receiving and visual interactive module, automatic sample preparation and rod system of carrier net;Sample intelligent delivery system, high-vacuum sample introduction system, including robot cooperation module, sample introduction collaborative module and sample introduction state machine control module;Intelligent electron microscope positioning imaging analysis system, including electron optical path self-adjusting module, sample nanometer level intelligent positioning module and image self-acquisition and real-time analysis module;State perception and intelligent scheduling system;Result output module and log record result storage module.Electron optical path self-adjusting module is built-in based on sensitivity matrix online identification astigmatism automatic correction, image self-acquisition and real-time analysis module is built-in based on image edge contrast monitoring automatic focusing.Intelligent closed loop from original sample preparation, sample transfer to result output is realized by state perception and intelligent scheduling.

Description

Technical Field

[0001] This invention relates to the field of intelligent analysis and detection technology of transmission electron microscopy, and in particular to a fully autonomous high-throughput intelligent transmission electron microscopy analysis system and method.

[0002] More specifically, the present invention relates to a fully autonomous high-throughput intelligent transmission electron microscopy analysis system and method that integrates intelligent sample transfer, high-vacuum sample introduction, autonomous adjustment of electronic optical path including automatic astigmatism correction based on online identification of sensitivity matrix, nanoscale intelligent sample positioning, autonomous image acquisition and real-time analysis including automatic focusing based on image edge contrast monitoring, state perception and intelligent scheduling, visual interaction, and result output. Background Technology

[0003] Transmission electron microscopy (TEM) is a high-resolution microscopic analytical instrument that uses a high-energy electron beam to penetrate a sample and image it through an electromagnetic lens. It can obtain structural information at the nanoscale and even atomic scale, and is widely used in fields such as energy catalysis materials, metallic materials, semiconductor devices, nanomaterials, life sciences, and biomedicine. For example, in energy catalysis, TEM can be used to analyze catalyst particles, clusters, single-atom sites, and interface structures; in semiconductors, it can be used to characterize device cross-sections, thin film defects, and interface element distribution; and in life sciences and medicine, it can be used to observe the ultrastructure of cells, virus particles, protein complexes, tissue sections, and pathological sections. Therefore, TEM is an important common tool supporting microstructural analysis, material property correlation, and biomedical characterization.

[0004] With the increasing demands for scientific research and industrial testing, transmission electron microscopy (TEM) analysis has shifted from manual observation of small samples to high-throughput, standardized, and quantitative analysis of multiple batches of samples. However, traditional TEM analysis workflows typically require manual completion of steps such as sample transfer, injection, region identification, electron path adjustment, image acquisition, image screening, image analysis, and result recording. This process is lengthy, involves many steps, and relies heavily on operator experience. Especially in scenarios requiring large-scale statistical analysis, batch sample comparison, pathological section screening, or quantitative analysis of catalyst particles, manual operation suffers from low efficiency, high subjectivity, insufficient repeatability, and limited data statistical scale, making it difficult to meet the demands of high-throughput intelligent analysis.

[0005] Existing technologies include some intelligent improvements to transmission electron microscopy (TEM) or microscopic analysis workflows, such as patented technologies like CN121601534A, which can achieve sample transfer, vacuum injection, sample stage motion control, automatic focusing, image acquisition, and image recognition and analysis. These technologies can reduce manual operation in certain stages, but most focus on single functional modules. There is still a lack of unified coordination between sample transfer, handover, high-vacuum injection, electronic optical path adjustment, target area positioning, image acquisition, real-time analysis, and result output. For TEM, the sample size is small, the posture requirements are high, and the injection process needs to cross the external environment and the high-vacuum environment of the electron microscope. Without continuous detection and interlocking control of the relationship between sample position, clamping status, sealing status, vacuum status, optical path status, imaging quality, and result recording, problems such as misordering, sample drop, abnormal injection, vacuum fluctuations, unstable image quality, or difficulty in locating abnormalities can easily occur.

[0006] For sensitive samples that are easily oxidized, deliquescent, or readily adsorb contaminants, or that require a high degree of cleanliness, the samples may be affected by air, water vapor, oxygen, particulate matter, or other contaminants during transfer from the sample area to the transmission electron microscope workstation, leading to changes in sample structure or distortion of characterization results. Conventional sample transport structures typically only consider mechanical handling and location access, making it difficult to provide temporary isolation, inert atmosphere, dry atmosphere, or vacuum protection before sample rod transfer, workstation docking, and high-vacuum sample injection.

[0007] In terms of electronic optical path adjustment, existing technologies have many shortcomings. For astigmatism correction, most existing automatic astigmatism correction technologies rely on precise pre-calibration before leaving the factory or before use. Once the working state of the electron microscope changes or environmental conditions fluctuate, the pre-calibrated parameters become invalid, requiring a cumbersome recalibration process. Some algorithms use fixed mapping relationships for compensation during the correction process, lacking online adaptability to the current actual state of the electron microscope, and the correction accuracy is limited by the nonlinear time-varying characteristics of the system. Some methods can achieve a certain degree of automatic adjustment, but their processes are incomplete, with defects in defocus state discrimination, contour feature extraction, or abnormal state handling, failing to guarantee stable operation under all working conditions. CN106062917A discloses an automatic astigmatism correction method, but this method suffers from insufficient astigmatism adjustment accuracy, slow convergence speed, and sensitivity to changes in the state of the electron microscope in practical applications, exhibiting poor robustness. CN112924477A proposes another astigmatism correction scheme. However, this scheme cannot cover the entire process from acquiring the original image to completing the correction. In some cases, it requires manual intervention or relies on pre-calibrated precision.

[0008] For autofocus, various approaches have been developed in the existing technology. One type of method estimates the magnitude and direction of defocus based on electron beam image shift, beam tilt, or related displacement, and then calculates the required current compensation value. Another type of method acquires multiple images at different focal positions and uses a sharpness evaluation function to calculate the optimal focal position that maximizes the sharpness index for each image. However, defocus estimation methods based on image shift or beam tilt often require additional hardware or specific electron beam scanning procedures, increasing system complexity and calibration difficulty. Methods based on multi-frame focal sequence scanning require acquiring multiple images to locate the optimal focal point, which significantly increases the electron irradiation dose to the sample and prolongs the focusing time. Furthermore, for electron beam-sensitive materials, excessively long irradiation times may lead to sample damage or structural changes. Furthermore, regardless of whether gradient, Laplacian, variance, or frequency domain energy is used as a sharpness metric, these metrics are all single scalars calculated for the entire image or a fixed region of interest. When there are contaminant particles, uneven background brightness, low signal-to-noise ratio, or sparse sample texture in the image, these global metrics are easily interfered with and lead to misjudgments. Most existing methods lack explicit self-detection logic for focus status and failure exit mechanisms in the automatic acquisition process. When the focusing process fails to converge due to sample drift, charging effect, or environmental disturbances, the system will continue to acquire images, wasting acquisition time and potentially rendering all subsequent data invalid due to improper focusing.

[0009] Therefore, there is an urgent need to provide a fully autonomous, high-throughput intelligent transmission electron microscopy (TEM) analysis system and method that can integrate functional modules such as intelligent sample transfer, high-vacuum sample introduction, autonomous adjustment of the electronic optical path, nanoscale intelligent sample positioning, autonomous image acquisition and real-time analysis, state perception and intelligent scheduling, visualization interaction, and result output. This would transform the TEM analysis process from a decentralized operation that relies on human experience into an intelligent analysis method with state perception, autonomous decision-making, closed-loop adjustment, real-time analysis, and process visualization capabilities. Summary of the Invention

[0010] In view of this, the present invention provides a fully autonomous high-throughput intelligent transmission electron microscopy analysis system and method to solve the problems of the disconnect between intelligent sample transfer, high-vacuum sample introduction, electronic optical path adjustment, sample positioning, image acquisition, image analysis, visualization interaction and result output in the existing transmission electron microscopy analysis process, which makes it difficult to achieve state perception, autonomous adjustment, closed-loop feedback, real-time analysis and anomaly interlocking.

[0011] To achieve the above objectives, this invention provides a fully autonomous high-throughput intelligent transmission electron microscope analysis system. Through a unified state perception and intelligent scheduling system, the sample flow, action flow, vacuum state flow, image analysis flow, and result recording flow are coupled, enabling the target sample to form an intelligent closed loop from sample transfer, robot handover, high-vacuum sample introduction, autonomous transmission electron microscope imaging and real-time intelligent analysis to result output under conditions of minimal human intervention.

[0012] The technical means employed in this invention are as follows: A fully autonomous high-throughput intelligent transmission electron microscope analysis system includes: The task receiving and visualization interaction module is used to receive the start command of the sample analysis process and call the system's preset sample analysis tasks, imaging parameters, sample injection process, image parsing strategy and safety interlock strategy. An automated sample preparation and rod loading system is used to perform quantitative processing, mesh supply and holding, quantitative dropping, drying, quality inspection, flexible and precise transfer of mesh and automatic rod loading on powder samples, liquid samples or dispersions to be added, according to the sample analysis task. After confirming the position and posture of the mesh and the effective status of the sample rod locking mechanism, the sample rod is updated to a ready-to-injection state for subsequent sample injection. The intelligent sample transfer system is used to transfer the target sample rod in the sample loading state to the transmission electron microscope workstation according to the sample analysis task, and to bring the target sample into the graspable position of the robot collaboration module. A high-vacuum sample introduction system is used to transfer the target sample from the external environment to the high-vacuum environment of the transmission electron microscope and maintain the stability of the vacuum state of the main cavity of the transmission electron microscope during the sample introduction process. The intelligent electron microscope positioning-imaging-analysis system is used to perform sample positioning, electronic optical path adjustment, image acquisition, and image analysis after the target sample enters the stage of the transmission electron microscope. The intelligent electron microscope positioning-imaging-analysis system includes an electronic optical path autonomous adjustment module, a sample nanoscale intelligent positioning module, and an image autonomous acquisition and real-time analysis module. The status perception and intelligent scheduling system is connected to the task receiving and visualization interaction module, the automatic sample preparation and rod loading system, the intelligent sample transfer system, the high vacuum sample introduction system, the intelligent electron microscope positioning-imaging-analysis system, the result output module, and the log recording and result storage module. It is used to uniformly schedule automatic sample preparation, rod loading, sample transfer, sample introduction, positioning, imaging, analysis, sample removal, and safety interlocks based on task start instructions, preset task parameters, sample preparation task status, rod loading final inspection results, composite robot status, workstation docking status, robot collaboration module status, sample introduction interface status, vacuum status, access control status, personnel detection status, emergency stop status, electronic optical path status, sample stage status, image acquisition status, image quality evaluation results, and image analysis results. The result output module is used to output image analysis results, quantitative statistical results, and analysis reports; The log recording and result storage module is used to record and store key operational data and analysis results online and in real time during task execution.

[0013] Furthermore, the electronic optical path autonomous adjustment module includes an automatic astigmatism correction unit. This unit acquires an electronic image, performs a fast Fourier transform on the image to obtain a spectrogram, extracts the contour of the Thon ring in the spectrogram using an image segmentation model, and determines whether the contour meets preset feature conditions. If not, a hill-climbing algorithm is used to adjust the focal length until the preset feature conditions are met. An ellipse fitting is performed on the contour that meets the preset feature conditions to obtain an elliptical curve, and the ellipticity and major axis angle of the current elliptical curve are calculated. Perturbations are applied to the X and Y axes of the electronic optical path astigmatism corrector, respectively, and the ellipticity is adjusted accordingly. The Hessian matrix and its inverse matrix are calculated based on the changes in quantity and the changes in hardware parameters of the X and Y axes. It is then determined whether the current ellipticity is less than the target threshold. If not, the X and Y axis adjustments of the astigmatism corrector are calculated using the inverse matrix and compensation is executed. After the electron microscope stabilizes, the new ellipticity vector after adjustment is obtained. It is then determined whether the adjusted ellipticity has deteriorated compared to the original. If it has deteriorated, coordinate rollback is performed and the gain multiplier is reduced before compensation is re-executed. If it has not deteriorated, the Hessian matrix is ​​updated using the Broyden method and iteration continues until the ellipticity is less than the target threshold.

[0014] Furthermore, the image autonomous acquisition and real-time analysis module includes an autofocus unit, which is used to acquire the current magnetic lens current value, the current magnification, and the current transmission electron microscopy image; perform model segmentation or threshold segmentation on the current transmission electron microscopy image to extract a mask image; perform sample edge extraction processing on the mask image to extract the sample edge contour; select multiple edge points at equal intervals on the sample edge contour; obtain the normal direction based on the contour curvature in the neighborhood of each edge point; align and average the grayscale sampling numbers in each normal direction using each edge point as a zero-point reference to obtain the relationship curve between the number and the average grayscale value; extract the number of peaks and the width of white edge pixels as focusing criteria parameters based on the relationship curve; determine whether the relationship curve meets the preset focusing state conditions, the preset focusing state conditions including that the relationship curve is single-peaked and the width of white edge pixels is within a specified range corresponding to the current magnification; if the conditions are met, the focusing is determined to be complete; if not, the change value of the magnetic lens current is calculated based on the number of peaks, the width of white edge pixels, and the current magnification, and the current adjustment is performed, iterating until the focusing is complete or the failure threshold is reached.

[0015] Furthermore, the intelligent sample delivery system includes: A composite robot transfer module, which is used to carry and transport the target sample; The workstation docking module is used to perform position detection, positioning, and error compensation on the composite robot transfer module. A sample protection and transport unit is used to protect the target sample during the transport process.

[0016] Furthermore, the composite robot delivery module includes a mobile chassis, a sample carrying unit, a sample locking unit, a navigation and positioning unit, an obstacle avoidance and perception unit, an environmental monitoring unit, and a communication unit; The sample carrying unit includes a sample rod receiving groove, at least two lateral limiting surfaces, and an axial unidirectional limiting part; the sample rod receiving groove is used to carry the sample rod; the lateral limiting surfaces are used to limit the lateral displacement and rotation of the sample rod about the vertical direction; the axial unidirectional limiting part is used to limit the unidirectional displacement of the sample rod along the axial direction, so that the sample rod is pushed to a predetermined axial position. The sample locking unit is used to restrict the displacement of the sample carrier relative to the mobile chassis during the operation of the mobile chassis, and to release the lock after the composite robot transfer module reaches the workstation docking module and completes the positioning.

[0017] Furthermore, the sample protection and transfer unit includes a temporary sample chamber disposed on the composite robot transfer module. The temporary sample chamber has a sample rod insertion port, a sample rod receiving cavity, and an openable and closable protective structure. When the composite robot transfer module acquires the sample rod, the sample rod is at least partially inserted into the temporary sample chamber to form dustproof, pollution-proof, or local isolation protection during sample transfer. The temporary sample chamber is also equipped with one or more of the following: a sealing gasket, an inert gas interface, a dry gas interface, a vacuum interface, a humidity detection unit, an oxygen content detection unit, or a pressure detection unit.

[0018] Furthermore, the high-vacuum sample introduction system includes: The robot collaboration module is used to grip the target sample or sample rod and perform sample transfer, coaxial alignment, propulsion, rotation, locking or retraction actions; the robot collaboration module includes a robot body, an end effector, a gripping detection unit, a force-displacement hybrid control unit, an attitude detection unit and a sample handover status confirmation unit; The sample introduction coordination module is used to work in conjunction with the sample introduction interface, sample introduction channel, sealing mating part and vacuum interface of the transmission electron microscope; The sample injection state machine control module is used to control the target sample to complete high-vacuum injection in stages based on the sample position, clamping state, coaxial alignment state, mechanical signal, displacement signal, torque signal, rotation angle signal, injection interface state, vacuum signal, and sample rod positioning signal.

[0019] Furthermore, the high-vacuum injection process controlled by the injection state machine control module includes at least one of the following stages: coaxial alignment, first injection, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second injection or torsion, stable connection judgment, third injection and arrival confirmation. The conditions for completing one advance are: the sample rod is advanced to a preset target position and the axial contact force between the sample rod and the injection port meets a preset threshold. The condition for the completion of the secondary propulsion or torsion is that the torque between the sample rod and the injection port meets the preset secondary threshold. The conditions for completing the three-stage advancement are: the sample rod is advanced to the preset three-stage target position and the axial contact force between the sample rod and the injection port meets the preset three-stage threshold.

[0020] Furthermore, the intelligent electron microscope positioning-imaging-analysis system includes: An electronic optical path autonomous adjustment module is used to perform spot centering, brightness adjustment, contrast adjustment, autofocus, and astigmatism adjustment based on electron microscope state parameters and image feedback. The electronic optical path autonomous adjustment module includes a spot centering unit, a brightness adjustment unit, a contrast adjustment unit, an autofocus unit, an astigmatism adjustment unit, and an optical path state evaluation unit. The sample nanometer-level intelligent positioning module is used to identify candidate imaging areas based on low-magnification preview images, high-magnification images, and sample stage coordinate information, and to position the target area to the imaging field of view; the sample nanometer-level intelligent positioning module includes a low-magnification search unit, a target area identification unit, a sample stage displacement control unit, an image registration unit, and a positioning error correction unit. The image autonomous acquisition and real-time analysis module is used to acquire transmission electron microscope images and simultaneously perform image quality evaluation, target recognition, image segmentation, quantitative statistics, and report generation. The image autonomous acquisition and real-time analysis module includes an imaging parameter calling unit, an image acquisition unit, an image quality evaluation unit, an image segmentation unit, a target recognition unit, a quantitative statistics unit, and a report generation unit.

[0021] Furthermore, the task receiving and visualization interaction module includes a startup operation unit, a task status display unit, a real-time image display unit, a real-time processing result display unit, an exception prompt unit, and a digital twin display unit; The startup operation unit is a one-key start button, a touch start button, a host computer startup command, or a remote startup command; The real-time image display unit is used to display real-time images from a transmission electron microscope. The real-time processing result display unit is used to display image quality evaluation results, target recognition results, segmentation results, quantitative statistical results, or report generation status; The digital twin display unit is used to display the virtual operating status of the composite robot, workstation docking module, robot collaboration module, high vacuum sample introduction system, sample stage, electronic optical path status, and image acquisition and analysis process based on system status signals.

[0022] Furthermore, the state perception and intelligent scheduling system includes a central control unit, a task scheduling unit, a safety interlocking unit, and an anomaly handling unit; The central control unit is used to communicate with each module and coordinate each module to perform sample analysis tasks; The task scheduling unit is used to generate or call task queues based on the availability status of the composite robot, the status of the robot collaboration module, the status of the high vacuum sample introduction system, the status of the transmission electron microscope, the image resolution status, and the task priority. The safety interlock unit is used to determine whether to allow the next action to be performed based on the composite robot's positioning status, robot gripping status, vacuum status, access control status, personnel detection status, emergency stop status, and sample injection status machine status. The anomaly handling unit is used to perform actions such as pausing, alarming, retrying, rolling back, transferring to manual confirmation, or generating maintenance tasks when problems such as misordering, sample drop, sample loss, robot jamming, sample injection timeout, vacuum anomaly, unqualified image quality, access control anomaly, composite robot obstacle avoidance anomaly, or communication anomaly occur.

[0023] On the other hand, this invention also discloses a fully autonomous high-throughput intelligent transmission electron microscopy analysis method, which is executed using the above-mentioned system and includes the following steps: S1: After receiving the start command, the state perception and intelligent scheduling system calls the system's preset sample analysis task, imaging parameters, sample introduction process, image analysis strategy and safety interlock strategy. S2: The state perception and intelligent scheduling system performs self-checks on each module. If the preset conditions are met, a task permission signal is generated; otherwise, it enters the abnormal hierarchical processing. S3: The state perception and intelligent scheduling system controls the composite robot transfer module to carry the target sample and move it to the transmission electron microscope workstation according to the task permission signal; for sensitive samples, the system controls the sample protection and transfer unit to enter the protection state; S4: After the composite robot transfer module arrives at the workstation, the workstation docking module performs position detection, positioning, and error compensation; after meeting the preset requirements, the sample locking state is released, and the target sample is adjusted to the graspable position of the robot collaboration module. S5: The robot collaboration module clamps the target sample, and after the clamping detection unit determines that the clamping is successful, it enters the high vacuum sample injection process; S6: The robot collaboration module moves the sample to the vicinity of the injection interface, and the injection collaboration module assists in completing coaxial alignment; S7: The sample injection state machine control module controls the sample to complete the high vacuum injection in stages according to the preset state machine logic, including coaxial alignment, first push, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second push or twist, stable connection judgment, third push and arrival confirmation. S8: After the high vacuum injection is completed, the sample nanoscale intelligent positioning module performs low-magnification search, candidate region identification, sample stage displacement control and image registration to position the target imaging area to the imaging field of view. S9: The electronic optical path autonomous adjustment module performs automatic astigmatism correction. Specifically, it acquires an electronic image and performs a fast Fourier transform to obtain a spectrum. It then uses an image segmentation model to extract the contour of the Thon ring in the spectrum and determines whether it meets preset feature conditions. If not, it uses a hill-climbing algorithm to adjust the focal length. For contours that meet the conditions, it performs elliptic fitting to obtain the ellipticity and major axis angle. It applies perturbations to the X and Y axes of the electronic astigmatism corrector, calculates the Hessian matrix and its inverse matrix based on the change in the ellipticity vector and the change in hardware parameters. It uses the inverse matrix to calculate the adjustment amount and performs compensation. It iteratively updates using the Broyden method until the ellipticity is less than the target threshold. Then, it executes... Automatic focusing: Acquire the current transmission electron microscope image, perform sample edge extraction processing on the image to extract the sample edge contour; select multiple edge points at equal intervals on the edge contour and calculate the normal direction; align and average the grayscale sampling numbers in each normal direction to obtain the relationship curve between the number and the average grayscale value; extract the number of peaks and the width of white edge pixels; determine whether the relationship curve is single-peaked and whether the width of white edge pixels is within the specified range corresponding to the current magnification; if satisfied, focus is considered complete; if not satisfied, calculate the change value of magnetic lens current based on the number of peaks, the width of white edge pixels, and the current magnification and perform adjustment, iterating until focus is complete; S10: When the state of the electronic optical path and the positioning state meet the preset requirements, the image autonomous acquisition and real-time analysis module acquires transmission electron microscope images and simultaneously performs image quality evaluation, image segmentation, target recognition, quantitative statistics and report generation. S11: The state perception and intelligent scheduling system determines whether the preset requirements are met based on the image quality evaluation results and image analysis results; if not, it controls readjustment or re-acquisition; if met, it confirms the results are valid; for multi-field or multi-region tasks, it repeatedly executes the sample positioning, electronic optical path adjustment, image acquisition and real-time analysis steps until the preset acquisition quantity or termination condition is reached. S12: The result output module outputs image analysis results, quantitative statistical results, and analysis reports; the log recording and result storage module stores key operational data and analysis results in real time; S13: After the analysis is completed, the high-vacuum sample introduction system executes the sample removal process; the robot collaboration module transfers the sample to the designated location; the composite robot transfer module transfers the sample to the designated area according to the scheduling instructions; S14: After the state perception and intelligent scheduling system confirms that the desampling is completed and all modules are reset, it generates a task completion status.

[0024] Further, the method specifically includes the following steps: S1: The operator issues a start command through the task receiving and visualization interaction module, or the start command is triggered by the host computer system, preset program or task queue; After receiving the start command, the status perception and intelligent scheduling system calls the system's preset sample analysis task, target imaging mode, target magnification, shooting area requirements, low dose strategy, number of images acquired, image analysis algorithm type, high vacuum sample introduction process, safety interlock conditions and abnormal handling strategy; S2: The state perception and intelligent scheduling system performs self-checks on the composite robot delivery module, workstation docking module, robot collaboration module, high vacuum sample introduction system, transmission electron microscope host, electronic optical path autonomous adjustment module, sample nanoscale intelligent positioning module, image autonomous acquisition and real-time analysis module, access control system, emergency stop system, and communication status; if the preset conditions are met, the system generates a task permission signal; otherwise, it enters the abnormal classification processing. S3: The state perception and intelligent scheduling system controls the composite robot transfer module to carry the target sample and move it to the transmission electron microscope workstation according to the task permission signal; for sensitive samples, the system controls the sample protection and transfer unit to enter the protection state, so that the target sample is in a temporary sample chamber, sealed accommodating cavity, inert atmosphere, dry atmosphere, low oxygen environment or vacuum environment during the composite robot transfer process. S4: After the composite robot transfer module arrives at the transmission electron microscope workstation, the workstation docking module performs position detection, positioning and error compensation on the composite robot transfer module; when the position signal, positioning signal and error compensation signal all meet the preset requirements, the system releases the sample locking state and adjusts the target sample to the graspable position of the robot collaboration module. S5: The robot collaboration module moves to a gripping position and clamps the target sample or sample rod; the clamping detection unit collects clamping force, clamping displacement, gripper positioning signal, pressure signal, visual feedback or contact status; if the clamping force is within the preset range and the gripper is in position, the clamping is determined to be successful, and the status perception and intelligent scheduling system allows entry into the high vacuum sample introduction process; otherwise, according to the anomaly type, the robot collaboration module is controlled to pause, retract, retry, transfer to the manual confirmation process, or enter the anomaly classification processing. S6: The robot collaboration module moves the sample rod or sample carrier to the vicinity of the transmission electron microscope sample inlet; the sample inlet coordination module assists the robot collaboration module in completing coaxial alignment based on the center positioning device, visual positioning signal, ranging signal, mechanical guidance signal, attitude detection signal, or sample inlet status signal; the status perception and intelligent scheduling system determines whether subsequent sample inlet actions are allowed based on the coaxial positioning signal, axial distance, angle deviation, clamping status, and sample inlet status. S7: The sample injection state machine control module controls the sample rod or sample carrier to complete the high vacuum injection in stages according to the preset state machine logic; the high vacuum injection includes coaxial alignment, first advancement, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second advancement or torsion, stable connection judgment, third advancement and arrival confirmation; if the detection signal corresponding to any state does not meet the preset conditions, the action exceeds the preset time, the number of retries exceeds the preset number, or triggers access control, emergency stop, vacuum abnormality, rodless detection abnormality, timeout blocking abnormality or personnel safety interlock, the sample injection state machine control module enters the abnormal handling state; S8: After the high vacuum injection is completed, the status perception and intelligent scheduling system reads the sample rod positioning signal, the sample stage status signal, the injection interface status signal and the vacuum stability signal; when the sample rod is in place, the sample stage status is normal, the vacuum status does not fluctuate abnormally and the safety interlock is not triggered, the sample nanoscale intelligent positioning module performs low-magnification search, candidate region identification, sample stage displacement control, image registration and positioning error correction to position the target imaging area to the imaging field of view; S9: The electronic optical path autonomous adjustment module adjusts one or more of the following based on the transmission electron microscope state parameters and image feedback: spot centering, brightness, contrast, autofocus, and astigmatism adjustment. The optical path state evaluation unit determines whether the optical path state meets the imaging preset. If not, the module performs the corresponding adjustment action according to the abnormality type and re-acquires the image for evaluation after adjustment. S10: When the state of the electronic optical path and the target area positioning state meet the preset requirements, the image autonomous acquisition and real-time analysis module acquires transmission electron microscope images according to the preset imaging parameters, and simultaneously performs image quality evaluation, image segmentation, target recognition, quantitative statistics and report generation. S11: The state perception and intelligent scheduling system determines whether the preset requirements are met based on the image quality evaluation results and image analysis results. If not, it controls the electronic optical path autonomous adjustment module to readjust according to the anomaly type, or controls the sample nanoscale intelligent positioning module to reselect the area, or controls the image acquisition unit to re-acquire. If the requirements are met, the system confirms that the image acquisition and analysis results are valid. For tasks requiring multi-field, multi-magnification, or multi-region acquisition, the sample positioning, electronic optical path adjustment, image acquisition, and real-time analysis steps are repeated until the preset acquisition quantity, coverage area, or termination condition is reached. S12: The result output module outputs image analysis results, quantitative statistical results, and analysis reports; the log recording and result storage module stores task logs, composite robot transfer logs, workstation docking logs, robot handover logs, high vacuum sample injection logs, electronic optical path adjustment logs, sample positioning logs, image acquisition logs, image quality evaluation results, image analysis results, abnormal alarm records, manual confirmation records, and report output records in real time, and associates them according to timestamps, task numbers, sample transfer order, sample injection status, image acquisition time, abnormal occurrence time, or result output time; S13: After the analysis is completed, the state perception and intelligent scheduling system controls the high vacuum sample introduction system to execute the sample removal process; the robot collaboration module transfers the sample to the composite robot transfer module, the sample recovery position, the abnormal buffer position, or the designated position; the composite robot transfer module transfers the sample to the designated area according to the scheduling instructions; S14: The status perception and intelligent scheduling system generates a task completion status after confirming that the sample return is complete, the sample recovery is complete, the robot collaboration module is reset, the high vacuum sample introduction system is reset, the transmission electron microscope is in normal condition, the image analysis process is completed, and the log recording is completed; the task receiving and visualization interaction module displays task completion information, analysis result summary, anomaly record summary, and equipment reset status.

[0025] Further, the single advance described in step S7 includes: a preset safe zone of 10-30 mm outside the injection port, and a preset contact zone of 0-10 mm outside the injection port; determining the advance speed of the sample rod based on the axial contact force between the sample rod and the injection port through admittance control; within the safe zone, the advance speed is constrained to a range of 5-20 mm / s; within the contact zone, the advance speed is constrained to a range of 0.5-5 mm / s.

[0026] Furthermore, the secondary propulsion or torsion mentioned in step S7 includes: determining the angle and direction of the primary torsion and the angle and direction of the secondary torsion based on the structure of the electron microscope inlet; torsioning the sample rod around the axis in the preset torsion direction by a preset angle, and detecting the torque in real time; if the torque and angle change curve successively shows the characteristics of rising, falling and plateau, and the torque meets the preset constraint, then the torsion is completed; if an abnormal torsion occurs, the torsion is withdrawn to the primary target position under the premise that the torsion does not exceed the safety threshold; if the torque exceeds the safety threshold, manual intervention is required.

[0027] Furthermore, the vacuum qualification confirmation in step S7 includes: determining whether the evacuation permit state is met within a preset time; if it is met, performing evacuation operation for a preset time; during the evacuation process, reading the vacuum indicator signal through the vacuum signal acquisition module; when the vacuum indicator signal enters the target state and remains stable within the preset time, or when the vacuum pressure drops to the target threshold and the rate of drop is within the preset range, it is determined that the qualified vacuum state has been reached.

[0028] Furthermore, the stable connectivity determination in step S7 includes: when the vacuum signal is within the target range, the torque is within the target range, and the position of the sample rod reaches the target tolerance with respect to the preset secondary target position, it is determined to be a stable connectivity.

[0029] Furthermore, the anomaly classification handling includes: for anomalies that do not affect the continued execution of the main process and can continue, the system records the anomaly information and continues to execute subsequent steps; for anomalies that can be recovered through reset, compensation, or retry, the system suspends the relevant actions and returns to the corresponding step or judgment node; for anomalies that require manual confirmation, the system outputs alarm information and waits for manual confirmation; for severe vacuum anomalies, emergency stop triggering, personnel entering dangerous areas, suspected sample drop or collision risk, and protection termination anomalies, the system stops the relevant actions and enters the protection state.

[0030] Compared with the prior art, the present invention has at least the following beneficial effects: First, this invention integrates task reception and visualization interaction, intelligent sample transfer, high-vacuum sample introduction, autonomous adjustment of the electronic optical path including automatic astigmatism correction based on online sensitivity matrix identification and automatic focusing based on image edge contrast monitoring, nanoscale intelligent sample positioning, autonomous image acquisition and real-time analysis, state perception and intelligent scheduling, result output, and log recording and result storage into a system. This enables the transmission electron microscopy (TEM) analysis process to possess state perception, autonomous adjustment, closed-loop feedback, real-time analysis, and process visualization capabilities. The sample transfer, handover, high-vacuum sample introduction, autonomous electron microscopy imaging, real-time intelligent analysis, and result output form an intelligent closed loop. Each step achieves state interlocking, signal feedback, and data correlation through the state perception and intelligent scheduling system.

[0031] Secondly, the present invention achieves intelligent transfer and high-precision handover of target samples between the external area and the transmission electron microscope workstation by combining a composite robot transfer module, a workstation docking module and a robot collaboration module, thereby reducing the risk of sample drop, loss, misordering and sample damage caused by manual transport.

[0032] Third, the present invention passively positions the sample rod by means of a sample rod bearing and limiting structure, uses a lateral limiting surface to limit the rotational deviation of the sample rod, and uses an axial unidirectional limiting part to determine the axial reference position of the sample rod, so that the sample rod maintains a stable posture during the transfer, docking and robot gripping process.

[0033] Fourth, the present invention provides protection against contamination, oxidation, deliquescence, or air exposure for sensitive samples through a sample protection and transport unit. The temporary sample chamber provides a local isolation environment for the sample rod during the transport process of the composite robot, and the sealed transfer sample rod can transfer the target sample in a safe atmosphere, dry atmosphere, low oxygen environment, or vacuum environment through a retractable protective structure, thereby reducing the risk of contamination or structural changes of sensitive samples before transport and injection.

[0034] Fifth, the present invention divides the sample introduction process into multiple detectable, identifiable, and reversible state segments through a high-vacuum sample introduction system and a sample introduction state machine control module, and sets corresponding detection signals, action permission conditions, and anomaly handling strategies in each segment, thereby improving the determinism, safety, and controllability of the high-vacuum sample introduction process.

[0035] Sixth, this invention utilizes the sample inlet channel, sealed fit structure, and vacuum transition process of the transmission electron microscope goniometer stage to safely transfer the sample from the external environment to the high vacuum environment of the electron microscope. It does not require limiting the solution to an independent external double-door vacuum transition chamber structure, which is beneficial to improving system compatibility and engineering adaptability.

[0036] Seventh, this invention employs an automatic astigmatism correction method based on online sensitivity matrix identification, eliminating the need for precise pre-calibration and enabling adaptive adaptation to changes in the electron microscope's state. By acquiring the spectrum through Fast Fourier Transform, the Thon ring contour is accurately extracted using the SAM3 image segmentation model and elliptic fitting is performed to construct an ellipticity vector containing both ellipticity and principal axis direction angles, achieving precise quantification of electron beam distortion. The Hessian matrix between the astigmatism corrector adjustment and the ellipticity vector is dynamically updated online using the Broyden quasi-Newton method. Closed-loop control gradually adjusts the astigmatism corrector's current value, causing the current ellipticity to gradually approach and converge to the target threshold. Coordinate rollback and gain annealing mechanisms prevent overshoot and oscillations during the correction process, resulting in advantages such as fast correction speed and strong robustness.

[0037] Eighth, this invention utilizes an autofocus method based on image edge contrast monitoring. By aligning and averaging the grayscale profile along the sample edge normal direction, the two-dimensional edge focusing problem is transformed into a one-dimensional stable curve analysis problem, significantly reducing the impact of random noise, isolated contaminants, and local abnormal textures on focus evaluation. By jointly using the number of curve peaks and the width of white edge pixels as focusing criteria, it can more accurately characterize the edge extension morphology in transmission electron microscopy than a single gradient, variance, or frequency domain energy index, improving the robustness of focus judgment. By establishing a current correction model associated with magnification, the current change value can be adaptively calculated based on the current magnification and edge morphology parameters, improving the consistency and convergence speed of autofocus at different magnifications. By setting a closed-loop focusing state detection process, the success or failure of focusing can be actively determined during automatic acquisition tasks, avoiding continuous image acquisition in a defocused state.

[0038] Ninth, this invention utilizes a nanometer-level intelligent sample positioning module to achieve automatic target area identification, sample stage displacement control, and positioning error correction, thereby improving target area positioning efficiency and sample analysis coverage. Furthermore, this invention employs an autonomous image acquisition and real-time analysis module to simultaneously execute image acquisition, image quality evaluation, image segmentation, target identification, quantitative statistics, and report output, thus improving the efficiency and statistical scale of transmission electron microscopy image analysis.

[0039] Tenth, the present invention displays the task status, equipment status, sample introduction status, real-time transmission electron microscope images, image quality evaluation results, and image analysis results in real time through the task receiving and visualization interaction module. It can also perform status mapping of the composite robot, robot collaboration module, high vacuum sample introduction system, and electron microscope imaging process through the digital twin display unit, thereby improving the visualization of the system operation process, the efficiency of anomaly location, and the convenience of manual intervention.

[0040] Eleventh, this invention records task logs, transmission logs, docking logs, high vacuum sample introduction logs, electronic optical path adjustment logs, sample positioning logs, image acquisition logs, quality evaluation results, image analysis results, and abnormal alarm records in real time during the task execution process, so that key states and results in the transmission electron microscopy analysis process can be queried and verified by timestamp. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1This is a schematic diagram of the overall layout of the fully autonomous high-throughput intelligent transmission electron microscope analysis system of the present invention.

[0043] Figure 2 This is a schematic diagram showing a partial interaction between the sample carrier and the robot end effector of the present invention.

[0044] Figure 3 This is a schematic diagram of the system architecture of the fully autonomous high-throughput intelligent transmission electron microscope analysis system of the present invention.

[0045] Figure 4 This is a flowchart illustrating the fully autonomous high-throughput intelligent transmission electron microscopy analysis method of the present invention.

[0046] Figure 5 This is a flowchart of the automatic correction method for transmission electron image astigmatism according to the present invention.

[0047] Figure 6 This is the original image of the sample with astigmatism in an embodiment of the present invention.

[0048] Figure 7 This is a fast Fourier transform preprocessed image of an image with astigmatism in an embodiment of the present invention.

[0049] Figure 8 This is a Fast Fourier Transform image during the correction process of an image with astigmatism in an embodiment of the present invention.

[0050] Figure 9 It is the corrected Fast Fourier Transform image corresponding to the astigmatic image in the embodiment of the present invention.

[0051] Figure 10 This is the original image of the sample after correction without astigmatism in the embodiments of the present invention.

[0052] Figure 11 This is a flowchart of the automatic focusing process of the transmission electron microscope of the present invention.

[0053] Figure 12 This is a mask image for extracting the edge of the sample in this invention.

[0054] Figure 13 This is a diagram illustrating edge extraction and edge point selection in this invention.

[0055] Figure 14 This is the normal generation diagram of the present invention.

[0056] Figure 15 This is a graph showing the alignment and average curve of the edge grayscale cross-section of the present invention. Detailed Implementation

[0057] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0058] Example 1 like Figure 1 and Figure 3 As shown, this embodiment provides a fully autonomous high-throughput intelligent transmission electron microscope analysis system, including a task receiving and visualization interaction module, an intelligent sample transfer system, a high-vacuum sample introduction system, an intelligent electron microscope positioning-imaging-analysis system, a status perception and intelligent scheduling system, a result output module, and a log recording and result storage module.

[0059] The task receiving and visualization interaction module is used to receive the start command of the sample analysis process and call the system's preset sample analysis tasks, imaging parameters, sample introduction process, image analysis strategy, and safety interlock strategy. In one embodiment, the task receiving and visualization interaction module includes a start operation unit, a task status display unit, a real-time image display unit, a real-time processing result display unit, an anomaly prompt unit, and a digital twin display unit. The start operation unit is one or more of a one-click start button, a touch start button, a host computer start command, or a remote start command; the real-time image display unit is used to display real-time images from a transmission electron microscope; the real-time processing result display unit is used to display image quality evaluation results, target recognition results, segmentation results, quantitative statistical results, or report generation status; the digital twin display unit is used to display the virtual operating status of the composite robot, workstation docking module, robot collaboration module, high-vacuum sample introduction system, sample stage, electron optical path status, and image acquisition and analysis process based on system status signals.

[0060] The sample protection and transport unit is used to protect the target sample during transport. The sample protection and transport unit includes one or more of the following: a temporary sample chamber, a protective cover, a sealed accommodating cavity, an inert atmosphere holding unit, a dry atmosphere holding unit, a vacuum holding unit, or a sealed transfer sample rod. The sample protection and transport unit reduces the risk of the target sample being affected by air, water vapor, oxygen, particulate matter, or other contaminants before being transferred by the composite robot, docked with the workstation, handed over by the robot, or introduced into a high-vacuum environment.

[0061] In this embodiment, the task receiving and visualization interaction module does not require the operator to manually input sample information, task parameters, or image analysis parameters. The system can pre-store or pre-configure sample analysis tasks, target imaging modes, magnification, imaging area requirements, low-dose strategies, number of images acquired, image analysis algorithm type, safety interlock conditions, and anomaly handling strategies. After the operator issues a start command through the task receiving and visualization interaction module, the status awareness and intelligent scheduling system automatically invokes the corresponding process based on the preset task and the current equipment status. In some implementations, the system can also extract target magnification, number of images acquired, image analysis strategies, low-dose strategies, or target area requirements based on built-in rules, historical detection schemes, literature search results, or externally pre-written process schemes, and autonomously generate or invoke detection processes accordingly. Literature search or historical scheme extraction can be completed before task start or pre-written into the system by an external system.

[0062] The intelligent sample transfer system is used to transfer target samples to a transmission electron microscope workstation according to the sample analysis task, and to place the target samples into the graspable position of the robot collaboration module. The intelligent sample transfer system may include a composite robot transfer module, a workstation docking module, and a sample protection and transport unit. The composite robot transfer module carries and transports the target sample; the workstation docking module performs position detection, positioning, and error compensation for the composite robot transfer module; and the sample protection and transport unit provides protection against contamination, oxidation, deliquescence, or air exposure during the transfer of the target sample.

[0063] The high-vacuum sample introduction system is used to transfer the target sample or sample carrier from the external environment to the high-vacuum environment of the transmission electron microscope (TEM) and maintain a stable vacuum state in the TEM's main cavity during the sample introduction process. The high-vacuum sample introduction system may include a robot collaboration module, a sample introduction coordination module, and a sample introduction state machine control module. The robot collaboration module is used to clamp the target sample or sample carrier and perform sample transfer, coaxial alignment, advancement, rotation, locking, unlocking, or retraction actions. The sample introduction coordination module works in conjunction with the TEM's sample introduction interface, sample introduction channel, sealing mating parts, and vacuum interface. The sample introduction state machine control module controls the target sample to complete the high-vacuum sample introduction in stages based on sample position, clamping state, coaxial alignment state, mechanical signals, displacement signals, torque signals, rotation angle signals, sample introduction interface state, vacuum signals, and sample rod positioning signals.

[0064] In one embodiment, the sample introduction coordination module is a goniometer stage coordination sample introduction module. This module works in conjunction with the sample inlet, sample channel, sealing mating part, and vacuum interface of the transmission electron microscope goniometer stage. After the robot coordination module clamps the sample rod, it aligns the sample rod coaxially with the goniometer stage's sample inlet. Under the control of the sample introduction state machine control module, it sequentially performs one advance, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, secondary advance or twisting, stable connection judgment, and tertiary advance, bringing the sample rod to the target sample introduction position. This embodiment does not imply that the invention must use a specific model or structure of goniometer stage; any structure capable of cooperating with the high-vacuum sample introduction interface of the transmission electron microscope to complete sample transfer can serve as the sample introduction coordination module of this invention.

[0065] The intelligent electron microscope positioning-imaging-analysis system is used to perform sample positioning, electronic optical path adjustment, image acquisition, and image analysis after the target sample enters the transmission electron microscope stage. The system includes an electronic optical path autonomous adjustment module, a sample nanoscale intelligent positioning module, and an image autonomous acquisition and real-time analysis module. The electronic optical path autonomous adjustment module performs spot centering, brightness adjustment, contrast adjustment, autofocus, and astigmatism adjustment based on the transmission electron microscope's state parameters and image feedback. The sample nanoscale intelligent positioning module identifies candidate imaging areas based on low-magnification preview images, high-magnification images, and stage coordinate information, and positions the target area within the imaging field of view. The image autonomous acquisition and real-time analysis module acquires transmission electron microscope images and simultaneously performs image quality evaluation, image segmentation, target recognition, quantitative statistics, and report generation.

[0066] The state perception and intelligent scheduling system is communicatively connected to the task receiving and visualization interaction module, the intelligent sample transfer system, the high-vacuum sample introduction system, the intelligent electron microscope positioning-imaging-analysis system, the result output module, and the log recording and result storage module. The state perception and intelligent scheduling system is used to uniformly schedule sample transfer, sample introduction, positioning, imaging, analysis, sample removal, and safety interlocks based on task start commands, preset task parameters, composite robot status, workstation docking status, robot collaboration module status, sample introduction interface status, vacuum status, access control status, personnel detection status, emergency stop status, electronic optical path status, sample stage status, image acquisition status, image quality evaluation results, and image analysis results.

[0067] The result output module is used to output image analysis results, quantitative statistical results, and analysis reports. The log recording and result storage module is used to record and store key operational data and analysis results online and in real time during task execution. The key operational data and analysis results include one or more of the following: task log, composite robot transfer log, workstation docking log, robot handover log, high-vacuum sample introduction log, electronic optical path adjustment log, sample positioning log, image acquisition log, image quality evaluation results, image analysis results, anomaly alarm records, manual confirmation records, and report output records. These records can be associated according to timestamps, task numbers, sample transfer order, sample introduction status, image acquisition time, anomaly occurrence time, or result output time.

[0068] Therefore, this embodiment does not simply set up the composite robot, the robot collaboration module, and the transmission electron microscope side by side. Instead, it enables the transmission electron microscope analysis process to have the capabilities of state perception, autonomous adjustment, closed-loop feedback, real-time analysis, visual interaction, and key process recording through state interlocking, signal feedback, and data association between task receiving and visual interaction, intelligent sample transfer, high-vacuum sample introduction, intelligent electron microscope positioning-imaging-analysis, state perception and intelligent scheduling, result output, and log recording and result storage.

[0069] This embodiment describes an intelligent sample transfer system. For example... Figure 1 and Figure 2 As shown, the intelligent sample transfer system may include a composite robot transfer module, a workstation docking module, and a sample protection and transport unit. The composite robot transfer module may include a mobile chassis, a sample carrying unit, a sample locking unit, a navigation and positioning unit, an obstacle avoidance sensing unit, an environmental monitoring unit, and a communication unit. The mobile chassis may be an AGV, AMR, wheeled mobile platform, track-based mobile platform, mobile robotic arm platform, linear module mobile platform, or other autonomous or semi-autonomous mobile mechanism capable of transferring samples between the sample location area and the transmission electron microscope workstation.

[0070] Example 2: Intelligent Sample Transfer and Protective Transportation This embodiment describes an intelligent sample transfer system. The intelligent sample transfer system may include a composite robot transfer module, a workstation docking module, and a sample protection and transport unit. The composite robot transfer module may include a mobile chassis, a sample carrying unit, a sample locking unit, a navigation and positioning unit, an obstacle avoidance sensing unit, an environmental monitoring unit, and a communication unit. The mobile chassis may be an AGV, AMR, wheeled mobile platform, track-based mobile platform, mobile robotic arm platform, linear module mobile platform, or other autonomous or semi-autonomous mobile mechanism capable of transferring samples between the sample area and the transmission electron microscope workstation.

[0071] The sample carrying unit is used to carry sample carriers, including sample boxes, sample holders, sample rods, sample trays, or other unmentioned sample carriers. The sample locking unit is used to restrict the displacement of the sample carrier relative to the mobile chassis during operation, and releases the lock after the composite robot transfer module reaches the workstation docking module and completes positioning. The navigation and positioning unit can employ laser navigation, visual navigation, QR code navigation, magnetic navigation, inertial navigation, SLAM navigation, preset path navigation, or a combination of multiple navigation methods. The obstacle avoidance and perception unit is used to detect personnel, obstacles, or abnormally approaching objects, and performs actions such as deceleration, stopping, detouring, replanning the path, or issuing an alarm when personnel or obstacles are detected entering a preset safe area.

[0072] In some embodiments, the sample carrying unit includes a sample rod carrying and limiting structure. The sample rod carrying and limiting structure may include a sample rod receiving groove, a first lateral limiting surface, a second lateral limiting surface, and an axial unidirectional limiting portion. After the sample rod is placed in the sample rod receiving groove, it can abut against the bottom support surface of the sample rod receiving groove, the first lateral limiting surface, or the second lateral limiting surface under the action of gravity. The first and second lateral limiting surfaces are used to limit the sample rod's offset in the horizontal plane and its rotational deviation about the vertical direction, ensuring that the sample rod maintains a predetermined posture after being transported by the composite robot transfer module to the workstation docking module.

[0073] The axial unidirectional limiting part is used to restrict the unidirectional displacement of the sample rod along the axial direction, so that the sample rod is pushed to a predetermined axial position. After the composite robot transfer module arrives at the workstation docking module, the sample rod can be pushed to the axial unidirectional limiting part by the robot collaboration module, sample transfer platform, pusher, or error compensation unit, so that the end of the sample rod abuts against the preset limiting position. Through this axial unidirectional limiting part, the axial reference of the sample rod in the sample carrying unit can be determined, thereby improving the repeatability of subsequent robot gripping, coaxial alignment, and high vacuum sample introduction.

[0074] In other embodiments, the sample rod receiving groove can be a V-shaped groove, U-shaped groove, arc-shaped groove, flat-bottomed groove, stepped groove, or an open limiting structure surrounded by multiple limiting blocks; the lateral limiting surface can be a plane, inclined surface, arc surface, conical surface, or a contoured surface adapted to the partial shape of the sample rod; the axial unidirectional limiting part can be an end block, positioning step, elastic limiting element, stop pin, limiting groove end face, or adjustable limiting block. All of the above structures can be used to achieve passive positioning without complex active clamping of the sample rod.

[0075] The workstation docking module is located at the transmission electron microscope workstation and is used for coarse and fine positioning of the composite robot transfer module, and to position the target sample within the grasping range of the robot collaboration module. The workstation docking module may include a robot docking position, a guiding structure, a positioning structure, a positioning detection unit, a sample transfer platform, and an error compensation unit. The guiding structure may be a guide groove, a guide rail, a V-shaped guide, a visual guidance mark, a magnetic navigation mark, a laser positioning mark, or a combination thereof; the positioning structure may be a positioning pin, a positioning hole, a limit block, an electromagnetic locking component, a clamping component, an adsorption positioning component, or a combination thereof; the error compensation unit may include a lifting platform, a lateral fine-tuning platform, a rotational fine-tuning platform, a visual servo compensation unit, a flexible guiding structure, or a combination thereof.

[0076] In some implementations, the intelligent sample transfer system also includes a sample protection and transport unit. For samples that are easily oxidized, deliquescent, easily adsorb contaminants, are susceptible to air exposure, or have high cleanliness requirements, the sample protection and transport unit provides environmental protection during the transfer of the sample rod from the sample area to the transmission electron microscope workstation.

[0077] In one embodiment, the sample protection and transport unit includes a temporary sample chamber disposed on the composite robot transfer module. The temporary sample chamber may include a chamber body, a sample rod insertion port, a sample rod receiving cavity, a limiting structure, and an openable / closable protective component. After the composite robot transfer module acquires the sample rod, the sample rod is at least partially inserted into the temporary sample chamber; the sample-bearing end of the sample rod is located in the sample rod receiving cavity, thereby reducing the time the sample-bearing end is directly exposed to the external environment during transport. The temporary sample chamber can be used to prevent sample contamination caused by particulate matter contamination, airflow disturbance, mechanical contact, or short-term air exposure.

[0078] In some implementations, the temporary sample chamber may also be equipped with one or more of the following: a sealing gasket, an elastic baffle, a sliding cover, a flip-up cover, an inert gas interface, a drying gas interface, a vacuum interface, a humidity detection unit, an oxygen content detection unit, or a pressure detection unit. The status perception and intelligent scheduling system can determine whether to allow the composite robot transfer module to continue the transfer operation or whether to allow the robot collaboration module to perform subsequent handover operations based on the status of the temporary sample chamber door, pressure, humidity, oxygen content, or the insertion position of the sample rod.

[0079] In another embodiment, the sample protection and transfer unit includes a sealed transfer sample rod. The sealed transfer sample rod includes a sample rod body, a sample carrier, a retractable protective sleeve, a sealing end cap, a seal, and a retaining cavity. The sample carrier is used to carry the sample to be analyzed; the retractable protective sleeve is fitted over the outside of the sample rod body and can move relative to the sample carrier between a protected position and an exposed position. When the retractable protective sleeve is in the protected position, it covers or surrounds the sample carrier, forming a sealed or semi-sealed protective cavity; when the retractable protective sleeve is in the exposed position, the sample carrier is exposed, allowing the sample to enter the target sample entry position or the imaging field of view of the transmission electron microscope.

[0080] The protective cavity of the sealed sample transfer rod can be an inert atmosphere cavity, a dry atmosphere cavity, a vacuum cavity, or a low-oxygen environment cavity. The sealing element can be an O-ring, a lip seal, an elastic gasket, a magnetic sealing structure, a metal sealing structure, or other sealing structures suitable for sample rod transfer. The sealed sample transfer rod may also be equipped with a protective sleeve position detection unit, a sealing status detection unit, a pressure detection unit, or an atmosphere detection unit to determine whether the sample is in a safe transfer state.

[0081] In other embodiments, the sample protection and transport unit may also be a sealed sample box, an inert atmosphere sample box, a vacuum transfer box, an openable protective cover, a local clean air curtain structure, a glove box docking structure, or a microenvironment transport chamber integrated with a composite robot transfer module. All of the above structures can be used to protect the target sample before sample transfer and high-vacuum injection.

[0082] Example 3: Robot Collaboration and Sample Handover like Figure 2 As shown, the robot collaboration module may include a robot body, an end effector, a gripping detection unit, a force-displacement hybrid control unit, an attitude detection unit, and a sample handover status confirmation unit. The robot body may be a multi-axis robotic arm, a collaborative robot, a Cartesian coordinate robot, a combination of linear modules and rotary mechanisms, a mobile robotic arm, or other actuators capable of transferring samples between the workstation docking module and the high-vacuum sample introduction system.

[0083] An end effector is used to clamp sample cassettes, sample holders, sample rods, sample trays, or other sample carriers. The end effector may include one or more of the following: mechanical grippers, vacuum suction ends, elastic clamping elements, limiting grooves, guide surfaces, flexible pads, anti-drop structures, sensor clamping mechanisms, or quick-change interfaces. For sample rod carriers, the end effector may be equipped with a semi-circular groove, V-groove, conical positioning part, clamping ring, or anti-rotation limiting structure adapted to the shape of the sample rod to improve the stability of the clamping posture.

[0084] The clamping detection unit is used to determine whether the target sample or sample carrier is reliably clamped. The clamping detection unit can make this judgment based on one or more of the following: clamping force, clamping displacement, gripper positioning signal, pressure signal, visual feedback, weight change, or contact state. When the clamping force is below a preset lower limit, the system determines that there is a risk of insufficient clamping or sample drop; when the clamping force is above a preset upper limit, the system determines that there is a risk of crushing damage or jamming; when the clamping displacement is inconsistent with the preset clamping displacement, the system determines that the sample carrier may be misaligned, tilted, or not in position.

[0085] The sample handover status confirmation unit is used to confirm whether the current sample meets the conditions for subsequent sample injection. This confirmation does not necessarily rely on a unique identifier carried by the sample itself, but can be based on the target sample's arrival status, the composite robot's transfer status, the robot's gripping record, the sample handover time, the task execution sequence, or the system operation log. In other embodiments, if the sample carrier is equipped with a QR code, barcode, RFID tag, or electronic tag, the sample handover status confirmation unit can also read this identifier and compare it with the task information for auxiliary verification.

[0086] The force-displacement hybrid control unit is used to jointly control the force and displacement feedback of the robot end effector when the robot collaborative module performs sampling, placement, docking, insertion, rotation, withdrawal, or retraction actions. The robot collaborative module can first perform a rapid approach action according to a preset displacement trajectory; after entering the low-speed alignment area or contact area, the force-displacement hybrid control unit reduces the movement speed based on axial force, radial force, torque, displacement, or rotation angle feedback, and uses compliant control, admittance control, impedance control, or force-limiting control methods to execute subsequent actions, thereby reducing the risk of rigid collisions between the sample carrier and the sample inlet, positioning structure, or other components.

[0087] If the robot collaboration module detects clamping failure, sample displacement, excessive posture deviation, abnormal contact force, motion timeout, personnel entering a dangerous area, or communication abnormalities with other modules during the handover process, the status perception and intelligent scheduling system controls the robot collaboration module to pause, retract, retry, issue an alarm, or initiate a manual confirmation process. Through the above methods, this embodiment can improve the reliability and safety of the sample handover process and provide a stable and clearly positioned sample carrier for subsequent high-vacuum sample injection.

[0088] Example 4: High Vacuum Sample Introduction System The high-vacuum sample introduction system includes a robot collaboration module, a sample introduction coordination module, and a sample introduction state machine control module. The high-vacuum sample introduction system is used to deliver the target sample or sample carrier from the external environment into the target sample introduction position of the transmission electron microscope, while maintaining the stability of the vacuum state in the main cavity of the transmission electron microscope during the process.

[0089] In one embodiment, the sample introduction coordination module is specifically a goniometer stage coordination sample introduction module. The goniometer stage coordination sample introduction module is used to work in conjunction with the sample inlet, sample introduction channel, sealing mating part, and vacuum interface of the transmission electron microscope goniometer stage. The goniometer stage coordination sample introduction module may include one or more of the following: a sample rod clamping and positioning unit, a center positioning device, a propulsion / rotation execution unit, a mechanical sensing unit, a vacuum signal acquisition unit, a goniometer stage status acquisition unit, a evacuation / pre-evacuation control unit, a rodless detection unit, and a timeout blocking detection unit.

[0090] The sample rod clamping and positioning unit is used to assist in positioning the sample rod's axial, radial, or angular orientation after it has been clamped by the robot collaborative module. The center positioning device ensures that the sample rod axis is coaxial or nearly coaxial with the axis of the goniometer inlet. The center positioning device can be a conical hole structure, a V-block structure, a spherical / conical mating structure, a guide sleeve structure, a visual positioning structure, a ranging positioning structure, a flexible guide structure, or other structures capable of achieving coaxial alignment between the sample rod and the inlet.

[0091] The propulsion / rotation actuator is used to drive the sample rod or sample carrier to move along the sample introduction direction, or to perform rotation, locking, unlocking, or torsion actions around its axis. The propulsion / rotation actuator can be implemented by the robot body itself, or by independent linear modules, rotary modules, motors, cylinders, lead screw mechanisms, clamping and rotating mechanisms, or combinations thereof. The mechanical sensing unit is used to collect axial force, radial force, contact force, torque, or other mechanical parameters during the sample introduction process. The vacuum signal acquisition unit is used to collect vacuum pressure, vacuum change rate, or vacuum qualification signals from the goniometer stage sample introduction channel, sample introduction interface, or related vacuum area.

[0092] The high-vacuum sample introduction process may include coaxial alignment, initial advancement, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, secondary advancement or twisting, stable connectivity judgment, tertiary advancement, and final confirmation. These stages can be added, removed, combined, or adjusted according to the actual transmission electron microscope structure and should not be construed as limitations on all implementation methods.

[0093] In other embodiments, the high-vacuum sample introduction system may also employ an independent pre-evacuation chamber, a vacuum transfer chamber, a vacuum sample exchange device, a mechanical sample transfer mechanism, or other sample introduction structures capable of transferring samples from the external environment to the high-vacuum environment. Regardless of the specific sample introduction structure used, as long as it divides the sample introduction process into multiple detectable, identifiable, and reversible stages based on multiple state signals, and forms an interlocked control with the state perception and intelligent scheduling system, it should be considered an equivalent or modified embodiment of the high-vacuum sample introduction system of the present invention.

[0094] Example 5: High Vacuum Sample Introduction State Machine Control In this embodiment, the sample introduction state machine control module controls the sample introduction process of the target sample from the external environment into the high vacuum environment of the transmission electron microscope in a state machine manner. The high vacuum sample introduction process is not a single continuous action, but is divided into multiple detectable, identifiable, and reversible state segments. Each state segment is equipped with input signals, action commands, completion conditions, timeout conditions, and anomaly handling strategies.

[0095] The state machine may include standby state, task confirmation state, composite robot arrival state, sample clamping confirmation state, coaxial alignment state, first propulsion state, sealing pre-tightening state, evacuation or pre-evacuation state, vacuum qualification confirmation state, second propulsion or torsion state, stable connection judgment state, third propulsion state, sample stage arrival confirmation state, imaging permission state, sample ejection state, sample recovery state, and anomaly handling state. These states can be added, removed, combined, or adjusted according to different electron microscope models, sample introduction interface structures, or sample carrier forms.

[0096] When the composite robot is in position, the composite robot transfer module enters the workstation docking module. The positioning detection unit, positioning structure, and error compensation unit output composite robot positioning signal, positioning completion signal, and error compensation completion signal, respectively. Only when all of the above signals meet the preset requirements does the state machine enter the sample clamping confirmation state.

[0097] In the sample clamping confirmation state, the robot collaboration module controls the end effector to clamp the target sample or sample rod. The clamping detection unit collects clamping force, clamping displacement, or jaw positioning signals. If the clamping force is within the preset range and the jaw is in position, the clamping is considered successful; if the clamping force is insufficient, the jaw is stuck, the sample is offset, or the clamping timeout is exceeded, the system enters an abnormal handling state, or performs re-clamping within a preset number of attempts.

[0098] In coaxial alignment, the robot collaboration module moves the sample rod or sample carrier to the vicinity of the injection port at a preset low speed, bringing the sample rod axis closer to the injection channel axis. The system calculates the attitude correction amount through visual positioning signals, ranging signals, attitude detection signals, feedback from the center positioning device, or mechanical limit signals, and controls the robot body to complete attitude compensation. The state machine enters a propulsion state only when the radial deviation, angular deviation, and axial distance are all within preset ranges.

[0099] In a single propulsion state, the robot collaboration module or propulsion / rotation actuator performs one propulsion along the sample injection axis. The conditions for completing one propulsion include the sample rod reaching a target position and the axial contact force meeting a threshold. If the target position is reached but the axial force is abnormal, or the axial force exceeds the safety limit, the state machine enters an abnormal handling state; if the threshold condition is not met, the propulsion can be repeated within a preset number of times.

[0100] In the pre-tightened sealing state, the robot collaboration module or propulsion / rotation actuator applies a preset axial pre-tightening force to the sample rod to form an initial seal between the sample rod and the sealing mating part at the sample inlet. Pre-tightening is completed when the axial contact force reaches a preset pre-tightening threshold and remains stable for a preset time. Insufficient pre-tightening force may indicate that a seal has not been established; excessive pre-tightening force may indicate hard contact or interference, causing the state machine to enter an abnormal processing state.

[0101] In evacuation or pre-evacuation mode, the evacuation / pre-evacuation control unit evacuates or pre-evacuates the sample inlet channel or related vacuum area. The vacuum signal acquisition unit reads the vacuum pressure, vacuum indication status, or vacuum descent rate. If the vacuum pressure enters the target range and remains stable within a preset time, or if the vacuum pressure descent rate is within a preset range, the system enters the vacuum qualification confirmation state; otherwise, it determines that the vacuum is abnormal and enters the abnormal handling state.

[0102] During secondary propulsion or torsion, the robot's collaborative module or propulsion / rotation actuator performs torsion, insertion, locking, unlocking, or secondary propulsion actions based on the current electron microscope sample introduction interface type. The state perception and intelligent scheduling system collects torque, rotation angle, axial displacement, and axial force in real time. The conditions for completing secondary propulsion or torsion include one or more of the following: torque entering a preset threshold range, rotation angle reaching the target angle, and sample rod position reaching the secondary target position.

[0103] In the stable connectivity assessment state, the system comprehensively determines whether the sample rod is stably connected to the sample introduction channel or the electron microscope vacuum system. Stable connectivity conditions include the vacuum signal being within the target range, the torque being within the target range, and the deviation between the sample rod position and the target position being within the target tolerance range, with these signals remaining stable for a preset time. If the stable connectivity conditions are met, the system enters a three-step advancement state; if not, it repeats the two-step advancement within a preset number of attempts or performs a retraction and retry.

[0104] During the three-stage propulsion phase, the robot's collaborative module or propulsion / rotation actuator performs the final propulsion at a low speed. The three-stage propulsion is completed when the sample rod reaches the target position three times and the axial contact force meets the threshold three times. If the target position is not reached during the three-stage propulsion or the axial force is abnormal, the three-stage propulsion will be repeated within a preset number of attempts; after exceeding the preset number of attempts, an error handling state will be entered.

[0105] In abnormal handling situations, the status awareness and intelligent scheduling system executes pause, rollback, retry, alarm, emergency stop, manual confirmation, or task termination based on the abnormality type. Abnormality types include one or more of the following: composite robot not in position, clamping failure, excessive coaxial deviation, primary propulsion limit failure, abnormal preload force, insufficient vacuum, abnormal secondary propulsion torque, stable connection failure, tertiary propulsion limit failure, rodless detection abnormality, sample injection timeout, access control abnormality, personnel entering a hazardous area, and emergency stop triggering.

[0106] Through the state machine control described above, this embodiment ensures that each segment of the high vacuum injection path has a clear detection signal, judgment condition, and anomaly handling logic, thereby improving the determinism, safety, and controllability of the high vacuum injection process.

[0107] Example 6: Intelligent Electron Microscope Positioning-Imaging-Analysis The intelligent electron microscope positioning-imaging-analysis system comprises an autonomous electronic optical path adjustment module, a nanometer-level intelligent sample positioning module, and an autonomous image acquisition and real-time analysis module. These three modules can form a closed-loop control relationship under the coordination of a state perception and intelligent scheduling system.

[0108] The electronic optical path autonomous adjustment module may include a spot centering unit, a brightness adjustment unit, a contrast adjustment unit, an autofocus unit, an astigmatism adjustment unit, and an optical path status evaluation unit. The optical path status evaluation unit determines whether the current electronic optical path status meets the imaging conditions based on one or more of the following: image grayscale distribution, brightness uniformity, edge sharpness, frequency domain energy distribution, directional texture, astigmatism ellipticity, or focal length response curve.

[0109] The sample nanoscale intelligent positioning module may include a low-magnification search unit, a target area recognition unit, a sample stage displacement control unit, an image registration unit, and a positioning error correction unit. Candidate imaging areas can be determined based on particle density, target particle size range, support film integrity, contamination area threshold, sample thickness, pore ratio, edge distance, lattice visibility, target area coverage, or user-defined selection criteria. The image registration unit establishes the mapping relationship between low-magnification image coordinates, high-magnification image coordinates, and sample stage coordinates. The positioning error correction unit corrects sample stage displacement errors based on the image registration results.

[0110] The image autonomous acquisition and real-time analysis module may include an imaging parameter calling unit, an image acquisition unit, an image quality evaluation unit, an image segmentation unit, a target recognition unit, a quantitative statistics unit, and a report generation unit. The image quality evaluation unit generates image usability evaluation results based on one or more of the following: image sharpness, signal-to-noise ratio, exposure status, drift degree, astigmatism degree, effective sample area ratio, contamination area, or sample damage characteristics. The image segmentation unit and target recognition unit are used to segment and identify particles, pores, defects, lattice regions, boundaries, or other target objects; the quantitative statistics unit outputs target quantity, size, area, perimeter, morphology factor, particle size distribution, spatial distribution, defect density, or other quantitative parameters; the report generation unit generates an analysis report based on the image analysis results.

[0111] When the image usability evaluation result does not meet the preset requirements, the state perception and intelligent scheduling system controls the electronic optical path autonomous adjustment module to perform readjustment, or controls the sample nanoscale intelligent positioning module to perform reselection, or controls the image acquisition unit to perform re-acquisition, based on the anomaly type. When the image usability evaluation result meets the preset requirements, the result output module outputs the image analysis results, quantitative statistical results, and analysis report, while the log recording and result storage module records the original image, processed image, quality evaluation results, analysis results, and report output records in real time.

[0112] This embodiment provides a detailed description of the automatic astigmatism correction unit in the electronic optical path autonomous adjustment module. like Figure 5 As shown, this invention provides an automatic correction method for transmission electron image astigmatism based on online sensitivity matrix identification, comprising the following steps: A1. Acquire an electronic image containing astigmatism using a transmission electron microscope detector. Crop the image into a square selection area with a size of 1024×1024 pixels. Perform a Fast Fourier Transform on the cropped image to generate a spectrogram.

[0113] Before performing the Fast Fourier Transform, the image is preprocessed to enhance its spectral quality. The preprocessing steps are as follows: interpolation is performed on the average pixel values ​​on both sides of the artifact to remove artifact interference; the average background brightness value is subtracted from the image and the median background noise is subtracted to eliminate background influence; the image brightness value is truncated to 15% to 99% and the truncated image is normalized to a grayscale value of 0 to 255.

[0114] After obtaining the spectrogram, the contour of the Thon ring is extracted using an image segmentation model. The image segmentation model used is a fine-tuned SAM3 model. The fine-tuning process is as follows: a large number of Fast Fourier Transform images are calibrated using the CVAT semi-automatic annotation tool according to preset rules or standards. The calibrated data is then used to retrain the open-source SAM3 weights, and the final fine-tuned weights are saved for contour recognition. The SAM3 model outputs a mask through text-guided segmentation, performs morphological closing operations on the mask to achieve contour closure, and then performs convex hull calculation to eliminate noise, ultimately obtaining a smooth and complete contour.

[0115] After extracting the contour, determine whether the extracted contour meets the preset feature conditions. The preset feature conditions are: mask value less than 4, solidity greater than 0.98, and area greater than 10000. If the contour does not meet these conditions, proceed to step S2; if it does, proceed to step S3.

[0116] A2. When the outline does not meet the preset feature conditions, it indicates that the current state is significantly out of focus, and the Thon ring in the spectrum may exhibit a hyperbolic shape. At this time, the focusing process based on the hill-climbing algorithm is triggered.

[0117] The hill-climbing algorithm works as follows: First, the focal length value is increased or decreased by one step to determine the search direction. The judgment criteria are to find a focal length value that meets three conditions: mask value less than 4, solidity greater than 0.98, and area greater than 10000, and the contour area reaches its maximum value. After finding this focal length value, it is set as the current focal length. Then, it returns to S1, reacquires the image, and performs contour extraction and judgment again until the contour meets the preset feature conditions.

[0118] A3. Perform ellipse fitting on the contours that meet the preset feature conditions. The fitting process is as follows: use the Sklansky algorithm to calculate the smallest convex polygon that encloses the contour, use this convex polygon to smooth the contour, and then use the least squares method to minimize the sum of the squares of the algebraic distances from all marked points to the ellipse curve. Connect the contour points to finally obtain the ellipse curve.

[0119] After obtaining the elliptic curve, calculate its ellipticity and major axis angle. Ellipticity is defined as the ratio of the major to the minor axis, and the major axis angle is defined as the angle between the major axis and the reference direction. Ellipticity and major axis angle together form a vector describing the current astigmatic state, where the major axis angle corresponds to the astigmatic direction and the ellipticity corresponds to the magnitude of the astigmatism.

[0120] A4. Apply known small perturbations to the X-axis and Y-axis of the electro-optical astigmatism corrector respectively. The perturbation method adopts the central difference scheme, that is, perform positive perturbation and negative perturbation respectively.

[0121] Record the changes in ellipticity and astigmatism direction before and after the disturbance, and also record the changes in the hardware parameters along the X and Y axes. Calculate the Hessian matrix and its inverse based on these two sets of changes. This Hessian matrix is ​​the sensitivity matrix, and the inverse matrix is ​​the compensation matrix. The calculation formula is as follows: The relationship between the change in the scatter vector and the change in hardware parameters is as follows:

[0122]

[0123]

[0124] in, The change in the astigmatism vector; , For the variation of hardware parameters of the image mirror, The second derivative of the change in the astigmatism hardware parameters with respect to the change in the astigmatism vector. For the sensitivity matrix, This is the compensation matrix.

[0125] A5. Determine whether the current ellipticity is less than the preset target threshold. The target threshold is set according to the actual application requirements of the electron microscope; in this embodiment, it is set to 1.15. If the ellipticity is less than the target threshold, astigmatism correction is completed, and the process ends; if the ellipticity is greater than or equal to the target threshold, then execute S6.

[0126] A6. Using the inverse matrix calculated in A4 and the current ellipticity vector, calculate the X-axis and Y-axis adjustments of the astigmatism corrector. The calculation method is to map the current ellipticity error to a hardware control variable change value through the inverse matrix, and then send this change value to the electromagnetic astigmatism corrector to perform compensation.

[0127] A7. After issuing the compensation command, wait for the electron microscope system to stabilize. The stabilization waiting time is set according to the electron microscope's response characteristics, typically ranging from several hundred milliseconds to several seconds. After stabilization, acquire new electron images again and calculate the adjusted new ellipticity vector following the procedures in S1 and S3.

[0128] S8. After obtaining the new ellipticity vector, compare the adjusted ellipticity with the original ellipticity. If the adjusted ellipticity is greater than the original ellipticity, it indicates that the compensation has deteriorated, and the coordinate rollback and gain annealing mechanism is triggered. Specifically, the following operations are performed: roll back to the coordinate position of the previous iteration point; halve the gain multiplier to reduce the risk of overshoot; and reset the integral term to eliminate accumulated errors. After completing the above operations, return to S6 to re-execute the compensation.

[0129] If the adjusted ellipticity is less than or equal to the unadjusted ellipticity, the compensation is considered effective. In this case, the Broyden method is used to update the Hessian matrix. The update method utilizes a quasi-Newton method to construct an iterative scheme, eliminating the need for further perturbation measurements in subsequent iterations. The iterative update scheme of the Broyden method is as follows:

[0130] A9. After completing the update in A8, check again whether the current ellipticity is less than the target threshold. If the ellipticity is less than the target threshold, astigmatism correction is complete, and the process ends. If the ellipticity is greater than or equal to the target threshold, return to A6 and repeat the loop of compensation calculation, application, state acquisition, and update judgment until the ellipticity converges to within the target threshold.

[0131] Specifically, first, an electronic image with significant astigmatism is acquired, and then the image is cropped to 1024×1024 pixels, such as... Figure 6 As shown, the cropped image undergoes Fast Fourier Transform preprocessing with image enhancement to obtain a spectrogram, as shown. Figure 7 As shown, the spectrum exhibits a distinct elliptical Thon ring.

[0132] The Thon ring contour was extracted using a fine-tuned SAM3 model. The extracted contour met the preset feature conditions of a mask value less than 4, a solidity greater than 0.98, and an area greater than 10000. Ellipse fitting was performed on the extracted contour, and the current ellipticity was calculated to be 1.945, which is greater than the set target threshold of 1.15.

[0133] Perturbations are applied to the X and Y axes of the electromagnetic astigmatism corrector, and the initial sensitivity matrix M and its inverse matrix M are calculated based on the changes in hardware parameters and astigmatism vector. - ¹.

[0134] Enter the core adjustment loop. Calculate the astigmatism adjustment amount using the inverse matrix and issue a compensation command. Obtain the new ellipticity after the electron microscope stabilizes. After the first compensation, the ellipticity is improved to 1.685. Figure 8 As shown, it is determined that no deterioration has occurred, and the Broyden method is used to update the Hessian matrix and proceed to the next iteration.

[0135] In a subsequent iteration, if the new ellipticity is greater than the ellipticity before compensation, the overshoot rollback and gain annealing mechanism is triggered, the previous image scatter point coordinates are returned, the gain multiplier is halved, and the iteration continues.

[0136] After several rounds of automatic cyclic compensation, the final spectrum is obtained, as shown below. Figure 9 As shown. At this point, the spectrum exhibits a perfectly circular and symmetrical distribution. The current ellipticity has converged to 1.10, which is less than the target threshold of 1.15, indicating that astigmatism adjustment has been successful. The final astigmatism-free original sample image is obtained, as shown. Figure 10 As shown, the entire automatic calibration process has ended.

[0137] This embodiment provides a detailed description of the autofocus unit in the image autonomous acquisition and real-time analysis module. like Figure 1 As shown, this invention provides an autofocusing method for transmission electron microscopy based on image edge contrast monitoring, comprising the following steps: B1. Obtain the current magnetic lens current value, current magnification, and current transmission electron microscopy image.

[0138] The magnetic lens current value is preferably the target magnetic lens current value that has a major influence on the current imaging focal plane, such as the objective lens current value, or an equivalent focusing control value formed by the combination of multiple lens currents. When the autofocus process starts, it first reads the current magnetic lens current value, current magnification, and the currently acquired transmission electron microscope image from the transmission electron microscope control system. After each magnetic lens current adjustment, it is necessary to wait for the image to stabilize before returning to S1 to reacquire the current magnetic lens current value, current magnification, and a new transmission electron microscope image to enter the next iteration. After each superposition of the magnetic lens current change value, image drift, global brightness fluctuation, and inter-frame differences can also be monitored. When the drift and brightness fluctuation are below the stabilization threshold, the image is determined to be stable, and the next round of focusing analysis is initiated.

[0139] B2. Perform model segmentation or grayscale conversion, threshold segmentation and morphological processing on the current transmission electron microscopy image to extract the mask image, and perform sample edge extraction processing on the mask image to extract the sample edge contour; Sample edge extraction can be achieved through a trained segmentation model or traditional image processing methods. The trained segmentation model can be a SAM3-based image segmentation model, used to receive transmission electron microscopy images and output a segmentation mask for the target sample region. After obtaining the segmentation mask, boundary extraction is performed to obtain the target sample edge contour. Traditional image processing methods may include threshold segmentation, connected component filtering, Canny edge detection, and contour extraction. Specifically, the transmission electron microscopy image can first be grayscaled and thresholded to obtain a binary image, which is then inverted to highlight dark sample regions. Subsequently, the contours of external connected regions are extracted, retaining the connected region with the largest area as the target sample region. Canny edge detection is then performed on the binary image of the target sample region, and contour extraction is performed on the detected edge image to obtain the target sample edge contour for focused analysis.

[0140] B3. Select multiple edge points at equal intervals on the sample edge contour of the current transmission electron microscopy image. Obtain the local tangent direction based on the contour curvature in the neighborhood of each edge point. Obtain the normal based on the tangent direction. Determine the background side based on the mask image. The direction from the inside of the sample to the background side outside the sample is the positive direction. B4. Using each edge point as a zero reference, align the grayscale sampling numbers in each normal direction to obtain multiple edge grayscale profiles, and average the multiple edge grayscale profiles to obtain the relationship curve between the number and the average grayscale value.

[0141] Each edge point is designated as a zero point, and grayscale sampling is performed along the normal direction of each edge point at fixed pixel or sub-pixel intervals. Before grayscale sampling, one or more preprocessing operations can be performed on the transmission electron microscope image, including noise reduction, normalization, adaptive filtering, histogram enhancement, or bandpass filtering. The grayscale sequences obtained from all edge points are aligned according to the zero points of each edge point, and all sequences are averaged to obtain an average grayscale curve. This average grayscale curve represents the relationship between the sequence number and the average grayscale value, and is used for subsequent extraction of focusing criterion parameters.

[0142] B5. Extract focus criteria parameters based on the relationship curve. The focus criteria parameters include the number of peaks and the width of white edge pixels.

[0143] Peak detection is performed on the average grayscale curve to obtain the number of peaks. Peak detection can be achieved through the average grayscale curve itself, the first derivative of the average grayscale curve, or the second derivative of the average grayscale curve. The average grayscale value at the zero point of each edge is used as the edge average grayscale. Sampling points in the average grayscale curve that are greater than the edge average grayscale value are identified as edge white pixels. The width of the white edge pixels is determined by the width of the continuous interval that meets this condition. The number of peaks is used to characterize the number of main peaks and sub-peaks of the edge grayscale average curve, and the width of the white edge pixels is used to characterize the width of the bright edge extension region at the sample edge. Both serve as quantitative criteria for determining the current focus state.

[0144] B6. Based on the focus criterion parameters, determine whether the relationship curve meets the preset focus state conditions. The preset focus state conditions include that the relationship curve is single-peaked and the width of the white edge pixels is within the specified range corresponding to the current magnification. If the conditions are met, determine that the current focus state is good and end the focus process.

[0145] A permissible width range corresponding to the current magnification is defined. This range is a threshold range associated with the magnification, pre-established by calibration experiments and stored in the form of a lookup table, empirical model, regression model, or neural network model. Different magnifications correspond to different target ranges for the white edge pixel width. When the number of peaks is 1 and the white edge pixel width falls within the permissible width range corresponding to the current magnification, the current image is considered to be in good focus, a focus completion signal is output, and the focusing process ends. If the number of peaks is not equal to 1 or the white edge pixel width exceeds the permissible range for the corresponding magnification, the current focus condition is not met, and the process proceeds to S7 for iterative correction.

[0146] B7. If the preset focus state conditions are not met, determine whether the current loop count is less than the threshold. If it is less than the threshold, calculate the change value of the magnetic lens current based on the number of peaks, the width of the white edge pixels, and the current magnification, and add the change value of the magnetic lens current to the current magnetic lens current value. After the image stabilizes, return to B1. If the current loop count is not less than the threshold, determine that the autofocus has failed and output an error message.

[0147] When the number of peaks is not equal to 1 or the width of the white edge pixels exceeds the allowable range of the corresponding magnification, it is determined whether the current iteration number is less than a preset threshold. If the current iteration number is greater than or equal to the threshold, autofocus is determined to have failed, an error message is output, and the process exits. If the current iteration number is less than the threshold, the change value of the magnetic lens current is calculated based on the number of peaks, the width of the white edge pixels, and the current magnification. The change value of the magnetic lens current is obtained through a pre-established correction mapping relationship, which takes at least the number of peaks, the width of the white edge pixels, and the current magnification as input, and outputs the current correction amplitude and correction direction. The correction direction is determined based on one or more of the following: the skewness of the average grayscale curve relative to zero, the positional relationship of the bright edge on the inside and outside of the sample, the adjustment direction of the previous iteration, and the pre-calibrated correction rules at the current magnification. The corrected magnetic lens current is updated to the sum of the current magnetic lens current value and the change value of the magnetic lens current. The transmission electron microscope is controlled to perform current adjustment, and after the image stabilizes, it re-enters S1 to continue iterating until the focusing state conditions are met or the failure threshold is reached.

[0148] The transmission electron microscope (TEM) autofocusing system based on image edge contrast monitoring, using the aforementioned method, includes an image acquisition module, an edge extraction module, an edge normal analysis module, a parameter extraction module, a focus determination module, a magnetic lens control module, and an anomaly handling module. The image acquisition module acquires the current magnetic lens current value, the current magnification, and the current TEM image. The edge extraction module performs sample edge extraction processing on the current TEM image to extract the sample edge contour. The edge normal analysis module selects multiple edge points at equal intervals on the sample edge contour, calculates the normal direction of each edge point, and defines the side where the gray value increases along the normal direction as positive. The parameter extraction module uses each edge point as a zero-point reference, aligns and averages the gray-level sampling numbers in each normal direction to obtain a relationship curve between the number and the average gray value, and extracts the number of peaks and the width of white edge pixels based on the relationship curve. The focus determination module determines whether the relationship curve meets preset focus conditions, which at least include a single-peak relationship curve and the width of white edge pixels within a specified range corresponding to the current magnification. The magnetic lens control module calculates the change in magnetic lens current based on the number of peaks, the width of white edge pixels, and the current magnification when the focus determination module determines that the preset focus conditions are not met, and then controls the magnetic lens to perform current adjustment. The exception handling module outputs an autofocus failure message when the current loop count reaches a threshold. All modules work together to execute the above methods.

[0149] Specifically, like Figure 11 As shown, this embodiment discloses an autofocusing method for transmission electron microscopy based on image edge contrast monitoring. This method operates within a closed-loop system consisting of transmission electron microscope control software, an edge recognition program, and a magnetic lens control unit, and includes the following steps: Step 101: Start the autofocus process and read the current magnetic lens current value I. t Current magnification M t And the currently acquired image I(x,y). The magnetic lens current value is preferably a target magnetic lens current value that has a major influence on the current imaging focal plane, such as the objective lens current value, or an equivalent focusing control quantity formed by a combination of multiple lens currents.

[0150] Step 102: Perform sample edge extraction processing on the current image I(x,y) to obtain the main sample region and / or sample edge contour C. The sample edge extraction processing can be implemented using various techniques: In a preferred embodiment, the current image can be input into the SAM3 model to obtain a sample main region segmentation mask S(x,y), and then boundary extraction can be performed on the target sample mask to obtain the sample edge contour C; In another optional embodiment, the Canny edge detection algorithm can be used to directly extract the sample edges, combined with edge connection, contour filtering, morphological repair, or region closure processing to obtain the target edge contour C for subsequent analysis; In other embodiments, threshold segmentation, active contour models, traditional image segmentation algorithms, or other basic visual models can also be used to achieve sample edge extraction.

[0151] When the sample edge extraction results contain multiple candidate regions or multiple candidate contours, the region or contour corresponding to the target sample can be selected as the object of subsequent focused analysis based on area, connectivity, aspect ratio, position prior, edge continuity, contour integrity, output confidence, or consistency with historical frames.

[0152] Step 103: Select N edge points p evenly along the sample edge contour C according to the arc length. i For each edge point p i Calculate the local tangent vector t based on its adjacent contour points. i And from this, the normal vector n is obtained. i Preferably, first in n i and -n i Take a set of neighboring sampling points in each of the two directions, and compare the average gray values ​​on both sides. If along n i If the direction grayscale value is higher, then n is retained. i As a positive sign; conversely, -n i Defined as positive, to ensure that the upward direction of grayscale values ​​is positive.

[0153] Step 104, for each edge point p i The corresponding position number is recorded as 0, and a grayscale sequence g of length 2K+1 is extracted along its normal direction at a set sampling interval Δs. i (k)=I(p i +k·Δs·n i ), where k takes values ​​in the range [-K, K]. All grayscale sequences are zero-aligned with k=0 and then averaged to obtain the average grayscale curve: g avg (k) = (1 / N) × Σ g i (k) In practice, Gaussian filtering, median filtering, Savitzky-Golay smoothing, or normalization can be applied to each grayscale sequence first, and then the average grayscale curve can be obtained to further suppress noise.

[0154] Step 105, based on the average grayscale curve g avg (k) Extract focusing parameters. Preferably, first extract the focusing parameters for g. avg (k) Perform peak detection and obtain the number of peaks P. t Then, the average gray value at the zero point of each edge point is taken as the edge average gray value T. e and will satisfy g avg (k)>T e The sampling points are identified as edge white pixels. For example, those satisfying g can be... avg (k)>T e The maximum width of a continuous interval is defined as the width of the white border W. t In other implementations, the number of consecutive sampling points that meet this condition can also be used as the white border width.

[0155] Step 106: Perform focus state determination. Preferably, set the magnification M... t The corresponding allowable width range [W] min (M t ), W max (M t When P t =1 and W t When the image falls within the allowed width range, the current image is determined to be in good focus, a focus completion signal is output, and the process ends.

[0156] Step 107, when P t Not equal to 1 or W t When the value exceeds the allowable range of the corresponding multiplier, check whether the current iteration number cnt is less than the preset threshold cnt. max If cnt is greater than or equal to cnt max If autofocus fails, an error message is output and the process exits; if cnt is less than cnt max Then proceed to the current correction step.

[0157] Step 108, according to P t W t and M t Calculate the change in magnetic lens current ΔI t Preferably, the correction function is established in advance through calibration: ΔI t = f(P t W t M t , σt ) Where, σ t The correction direction information can be obtained from the skew characteristics of the average grayscale curve relative to zero, the positional relationship of the bright edge on the inside and outside of the sample, the correction direction of the previous round, the empirical calibration rules of the corresponding magnification, or a combination thereof. When the direction does not need to be analyzed separately, the correction function can also directly output the signed current change value.

[0158] Step 109, update the corrected magnetic lens current to I. t+1 =I t +ΔI t The transmission electron microscope is controlled to perform current adjustment. After the image stabilizes, it is re-acquired and returned to step 101 to continue iterating until the focusing condition is met or the failure threshold is reached.

[0159] Through the above steps, this embodiment realizes an automatic focusing process based on changes in sample edge morphology, wherein sample edge extraction, parameter extraction, and focusing state detection are completed in the same closed loop.

[0160] Example 2: Edge Parameter Extraction Method In one alternative implementation, combined with Figures 12 to 15 The number of peaks P in step 105 t and white border width W t The following methods can be used to determine this: (1) such as Figure 12 As shown, region filtering or contour filtering is performed on the sample edge extraction results, preferably retaining the main region or main contour corresponding to the target sample, to obtain a mask or edge set for edge analysis; (2) For example Figure 13 As shown, the edge contour is extracted from the edge of the mask or the edge set, and multiple edge points are selected evenly according to the arc length; (3) such as Figure 14 As shown, normals are constructed at each edge point based on the local tangent vector, and the direction of grayscale increase is uniformly defined as positive. (4) such as Figure 15 As shown, the grayscale profiles in all normal directions are aligned and smoothed according to the zero point of the edge to obtain the average grayscale curve g. s (k); (5) The average gray value at the zero position of each edge point is taken as the edge average gray value T. e , will satisfy g s (k)>T e The sampling points are identified as edge white pixels, and the number of consecutive pixels or the width of a consecutive interval that meets this condition is defined as the white edge width W. t ; (6) For gs (k) Perform local maximum detection and retain only peaks whose height exceeds the threshold T. p Peak spacing exceeds threshold D p The peaks are obtained, and the number of peaks P is obtained. t ; (7) When P t =1 and W t When it falls within the normal range of the corresponding magnification, it is determined to be in focus; when P t Greater than 1, W t If the width is too wide or too narrow, it is judged as out of focus or abnormal focus.

[0161] In another alternative implementation, peak detection can also be performed on the first derivative curve or the second derivative curve to obtain a more sensitive edge diffusion criterion.

[0162] Alternatively, in an optional implementation, a correction mapping table for different magnifications can be established through offline calibration. Specifically, at several representative magnifications M, using the manually confirmed optimal focusing current as a reference point, the number of peaks, white edge width, and required correction current corresponding to different defocus states are recorded to form a lookup table or regression model.

[0163] For example, a piecewise linear model can be used: ΔI t = σ t × [ a(M t )×(W t -W0(M t )) + b(M t )×(P t -1) + c(M t ) ] Among them, W0(M t ) represents the target white border width corresponding to the optimal focus state at the current magnification, a(M t b(M) t ) and c(M t σ is the correlation coefficient obtained through calibration. t To correct the direction.

[0164] In another alternative implementation, the corrected mapping relationship can also be established through lightweight neural networks, support vector regression, random forests or other supervised learning models, thereby making the corrected current estimation more adaptive for different sample types.

[0165] Example 7: State Awareness, Visual Interaction, and Security Interlocking The state awareness and intelligent scheduling system may include a central control unit, a task scheduling unit, a safety interlock unit, an anomaly handling unit, and a visualization interaction unit. The central control unit communicates with each module and coordinates their execution of sample analysis tasks. The task scheduling unit generates or calls up task queues based on the availability of the composite robot, the status of the robot collaboration module, the high-vacuum sample introduction system, the transmission electron microscope status, the image resolution status, and task priorities. The safety interlock unit determines whether to allow the next action based on the composite robot's positioning status, robot gripping status, vacuum status, access control status, personnel detection status, emergency stop status, and sample introduction status machine status.

[0166] The state perception and intelligent scheduling system also transmits information such as the composite robot's position, workstation docking status, robot posture, clamping status, high-vacuum sample introduction status, sample rod position, vacuum status, sample stage position, electronic optical path status, image acquisition status, image analysis status, and abnormal alarm status to the task receiving and visualization interaction module for real-time display or digital twin display. This allows the operator to observe the system's operating status, real-time images, real-time processing results, and abnormal prompts in the visual interface, and to pause, confirm, reset, or manually take over as needed.

[0167] The safety interlock unit controls the relevant motion mechanisms to pause or enter a safe state when it detects personnel entering a dangerous area, abnormal access control opening, abnormal obstacle avoidance of the composite robot, robot gripping failure, abnormal sample inlet interface status, abnormal vacuum status, sample inlet timeout, abnormal rodless detection, or emergency stop signal triggering. The anomaly handling unit is used to execute pause, alarm, retry, rollback, transfer to manual confirmation, or generate maintenance tasks when problems such as out-of-order sampling, sample drop, sample leakage, robot jamming, sample inlet timeout, abnormal vacuum, unqualified image quality, abnormal access control, abnormal obstacle avoidance of the composite robot, or communication anomaly occur.

[0168] In one implementation, the exception handling unit can perform exception classification processing based on the degree of impact of the exception on the current task. For exceptions that do not affect the continuation of the main process and can continue, the system records the exception information and continues to execute subsequent steps; for exceptions that can be recovered through reset, compensation, or retry, the system pauses the relevant actions and returns to the corresponding step or judgment node; for exceptions that require manual confirmation, the system outputs alarm information and waits for manual confirmation; for protection termination exceptions such as severe vacuum anomalies, emergency stop triggering, personnel entering dangerous areas, suspected sample drop, or collision risk, the system stops the relevant actions and enters a protection state. The protection state may include one or more of the following: isolating the sample introduction channel, keeping the sample protection and transport unit in a protected state, controlling the robot collaboration module to retract, outputting alarm information, transferring to the manual confirmation process, or terminating the current task.

[0169] Example 8: Complete System Workflow like Figure 4 As shown, this embodiment provides a fully autonomous high-throughput intelligent transmission electron microscopy analysis method, which is executed using the aforementioned fully autonomous high-throughput intelligent transmission electron microscopy analysis system. The method may include the following steps.

[0170] Step S1, Task Initiation and Process Invocation. The operator issues a start command through the task receiving and visualization interaction module, or the start command is triggered by the host computer system, a preset program, or a task queue. After confirming the batch information, sample position information, sampling order, or task queue order of the samples to be analyzed, the status awareness and intelligent scheduling system, upon receiving the start command, invokes the system's preset sample analysis task, target imaging mode, target magnification, imaging area requirements, low-dose strategy, number of images acquired, image analysis algorithm type, high-vacuum sample introduction process, safety interlock conditions, and anomaly handling strategy. The sample analysis task and task parameters can be pre-stored within the system, or pre-written by an external system, or pre-configured by the program.

[0171] Step S2: System Self-Check and Task Authorization. The status awareness and intelligent scheduling system performs self-checks on the composite robot delivery module, workstation docking module, robot collaboration module, high-vacuum sample introduction system, transmission electron microscope main unit, electronic optical path autonomous adjustment module, sample nanoscale intelligent positioning module, image autonomous acquisition and real-time analysis module, access control system, emergency stop system, and communication status, and then executes the following steps: Figure 4 The J1 judgment determines whether the system status and safety interlocks meet preset requirements. If they do, the system generates a task permission signal; otherwise, it enters the anomaly classification process.

[0172] Step S3: Intelligent Sample Transfer and Protective Transport. The status perception and intelligent scheduling system controls the composite robot transfer module to carry the target sample and move it to the transmission electron microscope workstation based on the task permission signal. For sensitive samples, the system controls the sample protective transport unit to enter a protective state, ensuring the target sample is in a temporary sample chamber, sealed containment cavity, inert atmosphere, dry atmosphere, low-oxygen environment, or vacuum environment during the composite robot's transport process. The task receiving and visualization interaction module can simultaneously display the composite robot's position, operating path, transport status, or digital twin status.

[0173] Step S4, workstation docking. After the composite robot transfer module arrives at the transmission electron microscope workstation, the workstation docking module performs position detection, positioning, and error compensation on the composite robot transfer module. When the position signal, positioning signal, and error compensation signal all meet the preset requirements, the system releases the sample locking state and adjusts the target sample to a graspable position of the robot collaboration module.

[0174] Step S5: Robot handover and gripping confirmation. The robot collaboration module moves to a gripping position and grips the target sample or sample rod. The gripping detection unit collects gripping force, gripping displacement, gripper positioning signal, pressure signal, visual feedback or contact status, and executes... Figure 4 The system determines whether the transfer, docking, and clamping states meet preset requirements (J2). If they do, the state perception and intelligent scheduling system allows entry into the high-vacuum sample introduction process; if not, it controls the robot collaboration module to pause, revert, retry, switch to manual confirmation, or enter anomaly classification processing based on the anomaly type.

[0175] Step S6, High Vacuum Sample Injection Preparation. The robot collaboration module moves the sample rod or sample carrier to the vicinity of the transmission electron microscope's sample injection interface. The sample injection coordination module assists the robot collaboration module in completing coaxial alignment based on the center positioning device, visual positioning signals, ranging signals, mechanical guidance signals, attitude detection signals, or sample injection interface status signals. The status perception and intelligent scheduling system determines whether subsequent sample injection actions are permitted based on the coaxial positioning signals, axial distance, angular deviation, clamping status, and sample injection interface status.

[0176] Step S7, High Vacuum Sample Injection. The sample injection state machine control module controls the sample rod or sample carrier to complete the high vacuum sample injection according to the preset state machine logic. The high vacuum sample injection may include one or more of the following: coaxial alignment, single advance, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, secondary advance or torsion, stable connection judgment, tertiary advance and arrival confirmation. If the detection signal corresponding to any state does not meet the preset conditions, the action exceeds the preset time, the number of retries exceeds the preset number, or the access control, emergency stop, vacuum abnormality, rodless detection abnormality, timeout blocking abnormality, or personnel safety interlock is triggered, the sample injection state machine control module enters the abnormal handling state.

[0177] Step S8: Sample Positioning Confirmation and Intelligent Positioning. After high-vacuum sample introduction, the status sensing and intelligent scheduling system reads the sample rod positioning signal, sample stage status signal, sample introduction interface status signal, and vacuum stability signal. Only when the sample rod is in place, the sample stage status is normal, the vacuum status does not fluctuate abnormally, and the safety interlock is not triggered, does the system send an imaging permission command to the sample nanoscale intelligent positioning module. The sample nanoscale intelligent positioning module performs low-magnification search, candidate region identification, sample stage displacement control, image registration, and positioning error correction to position the target imaging area within the imaging field of view.

[0178] Step S9, Autonomous Adjustment of the Electron Optical Path. The autonomous adjustment module of the electron optical path autonomously adjusts one or more of the following based on the transmission electron microscope's status parameters and image feedback: spot centering, brightness, contrast, autofocus, and astigmatism adjustment. Figure 4The judgment J3 in the process determines whether the optical path state meets the imaging preset. If it does not, the electronic optical path autonomous adjustment module performs the corresponding adjustment action according to the anomaly type, and re-acquires the image for evaluation after adjustment.

[0179] Step S10: Autonomous Image Acquisition and Real-time Analysis. Once the electron optical path status and target area positioning status meet preset requirements, the autonomous image acquisition and real-time analysis module acquires transmission electron microscope (TEM) images according to preset imaging parameters, and simultaneously performs image quality assessment, image segmentation, target recognition, quantitative statistics, and report generation. The task receiving and visualization interaction module displays the TEM images, image quality assessment results, image segmentation results, target recognition results, quantitative statistics results, or analysis progress in real time.

[0180] Step S11, closed-loop judgment and repeated execution. The state awareness and intelligent scheduling system executes based on the image quality evaluation results and image parsing results. Figure 4 The system determines whether the image quality and resolution results meet preset requirements (J4). If the requirements are not met, the system controls the electronic optical path autonomous adjustment module to readjust, or controls the sample nanoscale intelligent positioning module to reselect the area, or controls the image acquisition unit to re-acquire the image, depending on the type of anomaly. If the requirements are met, the system confirms the validity of the image acquisition and resolution results. For tasks requiring multi-field, multi-magnification, or multi-region acquisition, the system can repeatedly execute the sample positioning, electronic optical path adjustment, image acquisition, and real-time resolution steps until the preset acquisition quantity, coverage area, or termination condition is reached.

[0181] Step S12: Result Output and Real-time Storage. The result output module outputs image analysis results, quantitative statistical results, and analysis reports. The log recording and result storage module stores the following logs in real time: task logs, composite robot transfer logs, workstation docking logs, robot handover logs, high-vacuum sample introduction logs, electronic optical path adjustment logs, sample positioning logs, image acquisition logs, image quality evaluation results, image analysis results, abnormal alarm records, manual confirmation records, and report output records. These records can be accurate to the second, millisecond, or other preset time precision, and can be correlated according to timestamps, task numbers, sample transfer order, sample introduction status, image acquisition time, abnormal occurrence time, or result output time.

[0182] Step S13, Sample Removal and Recovery. After analysis, the state perception and intelligent scheduling system controls the high-vacuum sample introduction system to execute the sample removal process. The sample removal process can be executed according to the reverse of the sample introduction or a preset sample removal state machine, and it detects the sample stage exit signal, sample rod position feedback, sample introduction interface status, vacuum pressure change, axial force, torque, rotation angle, composite robot positioning status, and robot clamping status. The robot collaboration module transfers the sample to the composite robot transfer module, sample recovery position, abnormal buffer position, or designated position. The composite robot transfer module transfers the sample to the designated area according to the scheduling instructions.

[0183] Step S14: Task Completion and Status Reset. The status awareness and intelligent scheduling system generates a task completion status after confirming that sample return is complete, sample recovery is complete, the robot collaboration module is reset, the high-vacuum sample introduction system is reset, the transmission electron microscope is functioning normally, the image analysis process is complete, and log recording is finished. The task receiving and visualization interaction module displays task completion information, a summary of analysis results, a summary of anomaly records, and the equipment reset status. The system then enters standby mode, waiting for the next startup command or the next preset task. In some implementations, multiple samples can be processed sequentially according to a preset task queue. The state awareness and intelligent scheduling system can perform sequential, parallel, or pipeline scheduling of the transfer, docking, injection, imaging, analysis, and retrieval of multiple samples based on the availability status of the composite robot, the availability status of the robot collaboration module, the status of the high-vacuum sample introduction system, the status of the transmission electron microscope, the image analysis status, task priority, and safety interlock status.

[0184] It should be understood that the order of the above steps can be adjusted according to the specific transmission electron microscope structure, sample type, sample carrying method, high-vacuum sample introduction method, and analytical task. For example, in some embodiments, the autonomous adjustment of the electron optical path can be pre-adjusted before sample positioning or fine-tuned after target area positioning; image quality evaluation can be performed immediately after each image acquisition or centrally after a set of image acquisitions; result output can be output in real time after each field of view acquisition or uniformly after all fields of view acquisitions are completed. The above adjustments do not affect the technical concept of this invention to achieve coordinated control of sample transfer, high-vacuum sample introduction, intelligent positioning imaging analysis, and result output through state perception and intelligent scheduling.

[0185] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fully autonomous high-throughput intelligent transmission electron microscope analysis system, characterized in that, include: The task receiving and visualization interaction module is used to receive the start command of the sample analysis process and call the system's preset sample analysis tasks, imaging parameters, sample injection process, image parsing strategy and safety interlock strategy. An automated sample preparation and rod loading system is used to perform quantitative processing, mesh supply and holding, quantitative dropping, drying, quality inspection, flexible and precise transfer of mesh and automatic rod loading on powder samples, liquid samples or dispersions to be added, according to the sample analysis task. After confirming the position and posture of the mesh and the effective status of the sample rod locking mechanism, the sample rod is updated to a ready-to-injection state for subsequent sample injection. The intelligent sample transfer system is used to transfer the target sample rod in the sample loading state to the transmission electron microscope workstation according to the sample analysis task, and to bring the target sample into the graspable position of the robot collaboration module. A high-vacuum sample introduction system is used to transfer the target sample from the external environment to the high-vacuum environment of the transmission electron microscope and maintain the stability of the vacuum state of the main cavity of the transmission electron microscope during the sample introduction process. The intelligent electron microscope positioning-imaging-analysis system is used to perform sample positioning, electronic optical path adjustment, image acquisition, and image analysis after the target sample enters the stage of the transmission electron microscope. The intelligent electron microscope positioning-imaging-analysis system includes an electronic optical path autonomous adjustment module, a sample nanoscale intelligent positioning module, and an image autonomous acquisition and real-time analysis module. The status perception and intelligent scheduling system is connected to the task receiving and visualization interaction module, the automatic sample preparation and rod loading system, the intelligent sample transfer system, the high vacuum sample introduction system, the intelligent electron microscope positioning-imaging-analysis system, the result output module, and the log recording and result storage module. It is used to uniformly schedule automatic sample preparation, rod loading, sample transfer, sample introduction, positioning, imaging, analysis, sample removal, and safety interlocks based on task start instructions, preset task parameters, sample preparation task status, rod loading final inspection results, composite robot status, workstation docking status, robot collaboration module status, sample introduction interface status, vacuum status, access control status, personnel detection status, emergency stop status, electronic optical path status, sample stage status, image acquisition status, image quality evaluation results, and image analysis results. The result output module is used to output image analysis results, quantitative statistical results, and analysis reports; The log recording and result storage module is used to record and store key operational data and analysis results online and in real time during task execution.

2. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The electronic optical path autonomous adjustment module includes an astigmatism automatic correction unit, which is used to acquire electronic images, perform fast Fourier transform on the electronic images to obtain a spectrum, and use an image segmentation model to extract the contour of the Thon ring in the spectrum and determine whether the contour meets the preset feature conditions. If the conditions are not met, the climbing algorithm is used to adjust the focal length until the preset feature conditions are met; the contour that meets the preset feature conditions is fitted with an ellipse to obtain an elliptical curve, and the ellipticity and major axis angle of the current elliptical curve are calculated. The X and Y axes of the electron microscope astigmatism corrector are perturbed respectively. The Hessian matrix and its inverse matrix are calculated based on the change in the ellipticity vector and the changes in the hardware parameters of the X and Y axes. It is determined whether the current ellipticity is less than the target threshold. If not, the adjustment amount of the X and Y axes of the astigmatism corrector is calculated using the inverse matrix and the compensation is executed. After the electron microscope stabilizes, the new ellipticity vector after adjustment is obtained. It is determined whether the ellipticity after adjustment has deteriorated compared with the original. If it has deteriorated, coordinate rollback is performed and the gain multiplier is reduced before compensation is re-executed. If it has not deteriorated, the Hessian matrix is ​​updated using the Broyden method and the iteration continues until the ellipticity is less than the target threshold.

3. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The image autonomous acquisition and real-time analysis module includes an autofocus unit, which is used to acquire the current magnetic lens current value, the current magnification, and the current transmission electron microscope image; perform model segmentation or threshold segmentation on the current transmission electron microscope image to extract a mask image; perform sample edge extraction processing on the mask image to extract the sample edge contour; select multiple edge points at equal intervals on the sample edge contour; obtain the normal direction based on the contour curvature in the neighborhood of each edge point; use each edge point as a zero-point reference; align and average the gray-level sampling numbers in each normal direction to obtain the relationship curve between the number and the average gray-level value. The number of peaks and the width of white edge pixels are extracted from the relationship curve as focusing criteria parameters; it is determined whether the relationship curve meets the preset focusing state conditions, which include the relationship curve being a single peak and the width of white edge pixels being within a specified range corresponding to the current magnification; if the conditions are met, focusing is determined to be complete; if not, the change value of magnetic lens current is calculated based on the number of peaks, the width of white edge pixels, and the current magnification, and current adjustment is performed, iterating until focusing is complete or the failure threshold is reached.

4. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The intelligent sample delivery system includes: A composite robot transfer module, which is used to carry and transport the target sample; The workstation docking module is used to perform position detection, positioning, and error compensation on the composite robot transfer module. A sample protection and transport unit is used to protect the target sample from the environment during the transport of the target sample. The composite robot delivery module includes a mobile chassis, a sample carrying unit, a sample locking unit, a navigation and positioning unit, an obstacle avoidance and perception unit, an environmental monitoring unit, and a communication unit. The sample carrying unit includes a sample rod receiving groove, at least two lateral limiting surfaces, and an axial unidirectional limiting part; the sample rod receiving groove is used to carry the sample rod; the lateral limiting surfaces are used to limit the lateral displacement and rotation of the sample rod about the vertical direction; the axial unidirectional limiting part is used to limit the unidirectional displacement of the sample rod along the axial direction, so that the sample rod is pushed to a predetermined axial position. The sample locking unit is used to limit the displacement of the sample carrier relative to the mobile chassis during the operation of the mobile chassis, and to release the lock after the composite robot transfer module reaches the workstation docking module and completes the positioning. The sample protection and transport unit includes a temporary sample compartment disposed on the composite robot transfer module. The temporary sample compartment has a sample rod insertion port, a sample rod receiving cavity, and an openable and closable protective structure. When the composite robot transfer module acquires the sample rod, the sample rod is at least partially inserted into the temporary sample compartment to form dustproof, pollution-proof, or local isolation protection during sample transport. The temporary sample chamber is also equipped with one or more of the following: a sealing gasket, an inert gas interface, a dry gas interface, a vacuum interface, a humidity detection unit, an oxygen content detection unit, or a pressure detection unit.

5. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The high-vacuum sample introduction system includes: The robot collaboration module is used to grip the target sample or sample rod and perform sample transfer, coaxial alignment, propulsion, rotation, locking or retraction actions; the robot collaboration module includes a robot body, an end effector, a gripping detection unit, a force-displacement hybrid control unit, an attitude detection unit and a sample handover status confirmation unit; The sample introduction coordination module is used to work in conjunction with the sample introduction interface, sample introduction channel, sealing mating part and vacuum interface of the transmission electron microscope; The sample injection state machine control module is used to control the target sample to complete high vacuum injection in stages according to the sample position, clamping state, coaxial alignment state, mechanical signal, displacement signal, torque signal, rotation angle signal, injection interface state, vacuum signal and sample rod positioning signal; The high-vacuum injection process controlled by the injection state machine control module includes at least one of the following stages: coaxial alignment, first injection, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second injection or twisting, stable connection judgment, third injection and arrival confirmation. The conditions for completing one advance are: the sample rod is advanced to a preset target position and the axial contact force between the sample rod and the injection port meets a preset threshold. The condition for the completion of the secondary propulsion or torsion is that the torque between the sample rod and the injection port meets the preset secondary threshold. The conditions for completing the three-stage advancement are: the sample rod is advanced to the preset three-stage target position and the axial contact force between the sample rod and the injection port meets the preset three-stage threshold.

6. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The intelligent electron microscope positioning-imaging-analysis system includes: An electronic optical path autonomous adjustment module is used to perform spot centering, brightness adjustment, contrast adjustment, autofocus, and astigmatism adjustment based on electron microscope state parameters and image feedback. The electronic optical path autonomous adjustment module includes a spot centering unit, a brightness adjustment unit, a contrast adjustment unit, an autofocus unit, an astigmatism adjustment unit, and an optical path state evaluation unit. The sample nanometer-level intelligent positioning module is used to identify candidate imaging areas based on low-magnification preview images, high-magnification images, and sample stage coordinate information, and to position the target area to the imaging field of view; the sample nanometer-level intelligent positioning module includes a low-magnification search unit, a target area identification unit, a sample stage displacement control unit, an image registration unit, and a positioning error correction unit. The image autonomous acquisition and real-time analysis module is used to acquire transmission electron microscope images and simultaneously perform image quality evaluation, target recognition, image segmentation, quantitative statistics, and report generation. The image autonomous acquisition and real-time analysis module includes an imaging parameter calling unit, an image acquisition unit, an image quality evaluation unit, an image segmentation unit, a target recognition unit, a quantitative statistics unit, and a report generation unit.

7. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The task receiving and visualization interaction module includes a startup operation unit, a task status display unit, a real-time image display unit, a real-time processing result display unit, an exception prompt unit, and a digital twin display unit. The startup operation unit is a one-key start button, a touch start button, a host computer startup command, or a remote startup command; The real-time image display unit is used to display real-time images from a transmission electron microscope. The real-time processing result display unit is used to display image quality evaluation results, target recognition results, segmentation results, quantitative statistical results, or report generation status; The digital twin display unit is used to display the virtual operating status of the composite robot, workstation docking module, robot collaboration module, high vacuum sample introduction system, sample stage, electronic optical path status, and image acquisition and analysis process based on system status signals.

8. The fully autonomous high-throughput intelligent transmission electron microscope analysis system according to claim 1, characterized in that, The state perception and intelligent scheduling system includes a central control unit, a task scheduling unit, a safety interlocking unit, and an anomaly handling unit. The central control unit is used to communicate with each module and coordinate each module to perform sample analysis tasks; The task scheduling unit is used to generate or call task queues based on the availability status of the composite robot, the status of the robot collaboration module, the status of the high vacuum sample introduction system, the status of the transmission electron microscope, the image resolution status, and the task priority. The safety interlock unit is used to determine whether to allow the next action to be performed based on the composite robot's positioning status, robot gripping status, vacuum status, access control status, personnel detection status, emergency stop status, and sample injection status machine status. The anomaly handling unit is used to perform actions such as pausing, alarming, retrying, rolling back, transferring to manual confirmation, or generating maintenance tasks when visual recognition failure, carrier net not being maintained, sample rod not being positioned, sample rod locking mechanism malfunction, rod loading failure, misordering, sample dropping, sample leakage, robot jamming, sample injection timeout, vacuum malfunction, image quality failure, access control malfunction, composite robot obstacle avoidance malfunction, or communication malfunction occur.

9. A fully autonomous high-throughput intelligent transmission electron microscopy analysis method, executed using the system of any one of claims 1 to 8, characterized in that, Includes the following steps: S1: After receiving the start command, the state perception and intelligent scheduling system calls the system's preset sample analysis task, imaging parameters, sample introduction process, image analysis strategy and safety interlock strategy. S2: The state perception and intelligent scheduling system performs self-checks on each module. If the preset conditions are met, a task permission signal is generated; otherwise, it enters the abnormal hierarchical processing. S3: The state perception and intelligent scheduling system controls the composite robot transfer module to carry the target sample and move it to the transmission electron microscope workstation according to the task permission signal; for sensitive samples, the system controls the sample protection and transfer unit to enter the protection state; S4: After the composite robot transfer module arrives at the workstation, the workstation docking module performs position detection, positioning, and error compensation; after meeting the preset requirements, the sample locking state is released, and the target sample is adjusted to the graspable position of the robot collaboration module. S5: The robot collaboration module clamps the target sample, and after the clamping detection unit determines that the clamping is successful, it enters the high vacuum sample injection process; S6: The robot collaboration module moves the sample to the vicinity of the injection interface, and the injection collaboration module assists in completing coaxial alignment; S7: The sample injection state machine control module controls the sample to complete the high vacuum injection in stages according to the preset state machine logic, including coaxial alignment, first push, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second push or twist, stable connection judgment, third push and arrival confirmation. S8: After the high vacuum injection is completed, the sample nanoscale intelligent positioning module performs low-magnification search, candidate region identification, sample stage displacement control and image registration to position the target imaging area to the imaging field of view; S9: The electronic optical path autonomous adjustment module performs automatic astigmatism correction. Specifically, it acquires an electronic image and performs a fast Fourier transform to obtain a spectrum. It then uses an image segmentation model to extract the contour of the Thon ring in the spectrum and determines whether it meets the preset feature conditions. If it does not meet the conditions, it uses a hill-climbing algorithm to adjust the focal length. It performs ellipse fitting on the contour that meets the conditions to obtain the ellipticity and major axis angle. It applies perturbations to the X-axis and Y-axis of the electronic astigmatism corrector, and calculates the Hessian matrix and its inverse matrix based on the change in the ellipticity vector and the change in the hardware parameters. The adjustment amount is calculated using the inverse matrix and compensation is performed; the Broyden method is used to iteratively update until the ellipticity is less than the target threshold; then automatic focusing is performed: the current transmission electron microscopy image is acquired, and the image is processed to extract the sample edge contour; multiple edge points are selected at equal intervals on the edge contour and the normal direction is calculated; the gray-level sampling numbers in each normal direction are aligned and averaged to obtain the relationship curve between the number and the average gray value. Extract the number of peaks and the width of white edge pixels; determine whether the relationship curve is single-peaked and whether the width of white edge pixels is within the specified range corresponding to the current magnification; if satisfied, determine that focusing is complete; if not satisfied, calculate the change value of magnetic lens current based on the number of peaks, the width of white edge pixels and the current magnification and perform adjustment, iterating until focusing is complete; S10: When the state of the electronic optical path and the positioning state meet the preset requirements, the image autonomous acquisition and real-time analysis module acquires transmission electron microscope images and simultaneously performs image quality evaluation, image segmentation, target recognition, quantitative statistics and report generation. S11: The state perception and intelligent scheduling system determines whether the preset requirements are met based on the image quality evaluation results and image parsing results; If the conditions are not met, the system will readjust or reacquire data; if the conditions are met, the results will be confirmed as valid. For multi-field or multi-region tasks, the sample positioning, electronic optical path adjustment, image acquisition, and real-time analysis steps will be repeated until the preset acquisition quantity or termination condition is reached. S12: The result output module outputs image analysis results, quantitative statistical results, and analysis reports; the log recording and result storage module stores key operational data and analysis results in real time; S13: After the analysis is completed, the high vacuum sample introduction system executes the sample removal procedure; The robot collaboration module transfers the sample to the designated location; the composite robot transfer module transfers the sample to the designated area according to the scheduling instructions; S14: After the state perception and intelligent scheduling system confirms that the desampling is completed and all modules are reset, it generates a task completion status.

10. The method according to claim 9, characterized in that, The method specifically includes the following steps: S1: The operator issues a start command through the task receiving and visualization interaction module, or the start command is triggered by the host computer system, preset program or task queue; after receiving the start command, the status perception and intelligent scheduling system calls the system's preset sample analysis task, target imaging mode, target magnification, shooting area requirements, low dose strategy, number of images acquired, image analysis algorithm type, high vacuum sample introduction process, safety interlock conditions and abnormal handling strategy; S2: The state perception and intelligent scheduling system performs self-checks on the composite robot delivery module, workstation docking module, robot collaboration module, high vacuum sample introduction system, transmission electron microscope host, electronic optical path autonomous adjustment module, sample nanoscale intelligent positioning module, image autonomous acquisition and real-time analysis module, access control system, emergency stop system, and communication status; if the preset conditions are met, the system generates a task permission signal; otherwise, it enters the abnormal classification processing. S3: The state perception and intelligent scheduling system controls the composite robot transfer module to carry the target sample and move it to the transmission electron microscope workstation according to the task permission signal; for sensitive samples, the system controls the sample protection and transfer unit to enter the protection state, so that the target sample is in a temporary sample chamber, sealed accommodating cavity, inert atmosphere, dry atmosphere, low oxygen environment or vacuum environment during the composite robot transfer process. S4: After the composite robot transfer module arrives at the transmission electron microscope workstation, the workstation docking module performs position detection, positioning and error compensation on the composite robot transfer module; when the position signal, positioning signal and error compensation signal all meet the preset requirements, the system releases the sample locking state and adjusts the target sample to the graspable position of the robot collaboration module. S5: The robot collaboration module moves to a gripping position and clamps the target sample or sample rod; the clamping detection unit collects clamping force, clamping displacement, gripper positioning signal, pressure signal, visual feedback or contact status; if the clamping force is within the preset range and the gripper is in position, the clamping is determined to be successful, and the status perception and intelligent scheduling system allows entry into the high vacuum sample introduction process; Otherwise, depending on the type of anomaly, the robot collaboration module will be paused, rolled back, retried, transferred to manual confirmation, or enter an anomaly classification process. S6: The robot collaboration module moves the sample rod or sample carrier to the vicinity of the transmission electron microscope sample inlet; the sample inlet coordination module assists the robot collaboration module in completing coaxial alignment based on the center positioning device, visual positioning signal, ranging signal, mechanical guidance signal, attitude detection signal, or sample inlet status signal; the status perception and intelligent scheduling system determines whether subsequent sample inlet actions are allowed based on the coaxial positioning signal, axial distance, angle deviation, clamping status, and sample inlet status. S7: The sample injection state machine control module controls the sample rod or sample carrier to complete the high vacuum injection in stages according to the preset state machine logic; the high vacuum injection includes coaxial alignment, first advancement, sealing pre-tightening, evacuation or pre-evacuation, vacuum qualification confirmation, second advancement or torsion, stable connection judgment, third advancement and arrival confirmation; if the detection signal corresponding to any state does not meet the preset conditions, the action exceeds the preset time, the number of retries exceeds the preset number, or triggers access control, emergency stop, vacuum abnormality, rodless detection abnormality, timeout blocking abnormality or personnel safety interlock, the sample injection state machine control module enters the abnormal handling state; S8: After the high vacuum injection is completed, the status perception and intelligent scheduling system reads the sample rod positioning signal, the sample stage status signal, the injection interface status signal and the vacuum stability signal; when the sample rod is in place, the sample stage status is normal, the vacuum status does not fluctuate abnormally and the safety interlock is not triggered, the sample nanoscale intelligent positioning module performs low-magnification search, candidate region identification, sample stage displacement control, image registration and positioning error correction to position the target imaging area to the imaging field of view; S9: The electronic optical path autonomous adjustment module adjusts one or more of the following based on the transmission electron microscope state parameters and image feedback: spot centering, brightness, contrast, autofocus, and astigmatism adjustment. The optical path state evaluation unit determines whether the optical path state meets the imaging preset. If not, the module performs the corresponding adjustment action according to the abnormality type and re-acquires the image for evaluation after adjustment. S10: When the state of the electronic optical path and the target area positioning state meet the preset requirements, the image autonomous acquisition and real-time analysis module acquires transmission electron microscope images according to the preset imaging parameters, and simultaneously performs image quality evaluation, image segmentation, target recognition, quantitative statistics and report generation. S11: The state perception and intelligent scheduling system determines whether the preset requirements are met based on the image quality evaluation results and image parsing results; If the conditions are not met, the electronic optical path autonomous adjustment module is controlled to readjust according to the type of anomaly, or the sample nanoscale intelligent positioning module is controlled to reselect the area, or the image acquisition unit is controlled to reacquire. If the conditions are met, the system confirms that the image acquisition and analysis results are valid. For tasks that require multi-field, multi-magnification, or multi-region acquisition, the sample positioning, electronic optical path adjustment, image acquisition, and real-time analysis steps are repeated until the preset acquisition quantity, coverage area, or termination condition is reached. S12: The result output module outputs image analysis results, quantitative statistical results, and analysis reports; the log recording and result storage module stores task logs, composite robot transfer logs, workstation docking logs, robot handover logs, high vacuum sample injection logs, electronic optical path adjustment logs, sample positioning logs, image acquisition logs, image quality evaluation results, image analysis results, abnormal alarm records, manual confirmation records, and report output records in real time, and associates them according to timestamps, task numbers, sample transfer order, sample injection status, image acquisition time, abnormal occurrence time, or result output time; S13: After the analysis is completed, the state perception and intelligent scheduling system controls the high vacuum sample introduction system to execute the sample removal process; the robot collaboration module transfers the sample to the composite robot transfer module, the sample recovery position, the abnormal buffer position, or the designated position; the composite robot transfer module transfers the sample to the designated area according to the scheduling instructions; S14: The state perception and intelligent scheduling system generates a task completion status after confirming that sample ejection is complete, sample recovery is complete, the robot collaboration module is reset, the high vacuum sample introduction system is reset, the transmission electron microscope is in normal status, the image analysis process is completed, and the log recording is completed; the task receiving and visualization interaction module displays task completion information, analysis result summary, anomaly record summary, and equipment reset status. The secondary propulsion or torsion mentioned in step S7 includes: determining the angle and direction of the primary torsion and the angle and direction of the secondary torsion based on the structure of the electron microscope inlet; torsioning the sample rod around the axis in the preset torsion direction by a preset angle, and detecting the torque in real time; if the torque and angle change curve successively shows the characteristics of rising, falling and plateau, and the torque meets the preset constraint, then the torsion is completed; if an abnormal torsion occurs, the torsion is withdrawn to the primary target position under the premise that the torsion does not exceed the safety threshold; if the torque exceeds the safety threshold, manual intervention is required. The vacuum qualification confirmation in step S7 includes: determining whether the evacuation permit state is met within a preset time; if it is met, performing evacuation operation for a preset time; during the evacuation process, reading the vacuum indicator signal through the vacuum signal acquisition module; when the vacuum indicator signal enters the target state and remains stable within the preset time, or when the vacuum pressure drops to the target threshold and the rate of drop is within the preset range, it is determined that the qualified vacuum state has been reached. The stable connectivity determination in step S7 includes: when the vacuum signal is within the target range, the torque is within the target range, and the position of the sample rod reaches the target tolerance with respect to the preset secondary target position, it is determined to be a stable connectivity. The anomaly classification and handling includes: for anomalies that do not affect the continued execution of the main process and can continue, the system records the anomaly information and continues to execute subsequent steps; for anomalies that can be recovered through reset, compensation, or retry, the system suspends the relevant actions and returns to the corresponding step or judgment node; for anomalies that require manual confirmation, the system outputs alarm information and waits for manual confirmation; for severe vacuum anomalies, emergency stop triggering, personnel entering dangerous areas, suspected sample drop or collision risk, and protection termination anomalies, the system stops the relevant actions and enters protection mode.

Images