Sample table automatic positioning analysis cavity based on visual feedback and sample transferring method

Abstract

The invention discloses a sample table automatic positioning analysis cavity based on visual feedback and a sample transferring method. The mechanical structure comprises an upper cavity, a lower cavity and a first communicating cavity communicating the upper cavity with the lower cavity and provided with a gate valve. The upper end of the upper cavity is provided with a composite motion device integrated with X-axis, Y-axis and Z-axis displacement tables and a polar angle rotating device through a second communication cavity; the side wall of the lower cavity communicates with the analyzer. Aiming at the problem of depth information loss existing in monocular vision, the system adopts a binocular vision system to determine three-dimensional space coordinates (X, Y and Z) of visual marks on a sample table, so that the problem of depth blur is fundamentally solved, and a core guarantee is provided for realizing micron-sized positioning. The processing module pre-stores standard positioning coordinates and integrates image acquisition, information processing and position control functions. And micron-order positioning precision, intelligent safety guarantee, thermal deformation real-time compensation and full-process automation are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of synchrotron radiation experimental equipment, and particularly relates to a sample table automatic positioning analysis cavity based on visual feedback and a sample transmission method. BACKGROUND

[0002] The ARPES (Angle-Resolved Photoemission Spectroscopy) experiment, which utilizes the high brightness, high collimation and adjustable polarization of the synchrotron radiation source, has become an ultra-microscope for revealing the microstructure and dynamic process of matter, and is widely applied in the fields of material science, physics, chemistry and biology. The high-resolution imaging, surface analysis and spectroscopy measurement of the ARPES experiment have strict requirements on the sample environment: the experiment is usually carried out in a vacuum or ultra-high vacuum state to eliminate the interference of gas molecules on the sample surface and the signal (such as photoelectrons) and ensure the measurement accuracy.

[0003] The ARPES experimental equipment mainly includes a sample inlet cavity for placing a sample holder to be measured from an atmospheric environment and initially pumping the sample holder to a high vacuum state. A vacuum transfer cavity is connected with the sample inlet cavity and the analysis cavity and serves as a transfer station for the sample holder. The analysis cavity is an ultra-high vacuum main experimental cavity for ARPES measurement. In addition, there are a sample processing cavity, a spin detector and an energy analyzer.

[0004] The analysis cavity, as one of the core components of the experimental station, has an upper cavity and a lower cavity inside, and a sample table is responsible for carrying, transferring and accurately positioning the sample in the vacuum environment. The performance of the sample table is directly related to the experimental safety and data quality. Taking the ARPES experiment as an example, the spot of the fourth-generation synchrotron radiation device can reach 3-5 microns, which poses an unprecedented challenge to the positioning accuracy of the sample table. When the sample and the detector (hemispherical electron energy analyzer) are deviated by more than this range, the irradiation area of the synchrotron radiation will deviate from the target area, and the signal detected by the detector will also be significantly attenuated.

[0005] The sample table system in the prior art usually adopts an open-loop or semi-open-loop mechanical driving mode, such as driving a micrometer through a stepper motor for displacement. Such a system mainly has the following technical defects: High collision risk: when the sample is transferred from one vacuum cavity to another vacuum cavity or approaches a precision instrument such as a detector, the main reliance is on the visual observation and manual operation of the operator. Due to the limited observation angle, visual blind area and possible human negligence, mechanical collision accidents between the sample table and its attached components and the key equipment such as the detector often occur, causing significant damage to the equipment and interruption of the experiment; Low positioning accuracy: the repeatability of the traditional mechanical structure is usually at the level of 10 microns, which is difficult to meet the matching requirements of the micron-level spot of the modern synchrotron radiation, resulting in a decrease in the signal acquisition quality; Thermal deformation has a great impact: when the temperature is changed (for example, from 10K to 300K), the sample rod will be displaced by millimeters due to thermal expansion and contraction, causing the light spot to deviate from the target measurement area, and the position consistency of long-time measurement cannot be guaranteed. The existing system lacks the ability to compensate for such dynamic drift in real time. Low degree of automation: the transfer and positioning process of the sample is heavily dependent on manual operation and intervention, which is low in efficiency and poor in reliability, and is difficult to meet the modern experimental demand of high throughput and high efficiency. SUMMARY

[0006] The purpose of the present application is to provide a sample stage automatic positioning analysis cavity based on visual feedback and a sample transfer method, which aims to solve the technical problems of low positioning accuracy, high collision risk, inability to compensate for thermal deformation and low degree of automation in the prior art.

[0007] To solve the above technical problems, the present application is realized by the following technical scheme: A sample stage automatic positioning analysis cavity based on visual feedback, the analysis cavity has an upper cavity and a lower cavity inside; The lower end of the upper cavity is connected to the lower cavity through a first communication cavity, and the first communication cavity is provided with a gate valve, and the upper end is installed with a composite motion device through a second communication cavity; The side wall of the lower cavity is connected to an analyzer; The output end of the composite motion device is connected to the sample stage through a sample rod, for controlling the sample stage to adjust in multiple degrees of freedom; The upper cavity and the lower cavity correspond to optical windows and are respectively installed with first and second binocular vision systems; The processing module is integrated with an image acquisition unit, an information processing unit and a position control unit; The image acquisition unit is used to drive the first and second binocular vision systems to acquire real-time images of the visual marks of the sample stage; The information processing unit is used to pre-process the images and convert the image data into three-dimensional coordinates and pose of the sample stage based on the binocular stereo vision principle; The position control unit is used to adjust the sample stage in multiple degrees of freedom through the composite motion device.

[0008] Further, the composite motion device includes an X-axis horizontal displacement table, a Y-axis horizontal displacement table, a Z-axis vertical displacement table and a polar angle rotating device in the second communication cavity, and each is independently provided with a stepping motor for driving, for three-dimensional coordinate correction, to ensure that the sample is always in the position of the rotation center during rotation.

[0009] Further, the first binocular vision system comprises a first camera and a second camera, and is configured to synchronously collect position information of the sample table in the upper cavity through the first camera and the second camera; the second binocular vision system comprises a third camera and a fourth camera, and is configured to jointly solve the position information of the sample table in the lower cavity and the three-dimensional spatial position information relative to the analyzer.

[0010] A sample table automatic positioning analysis cavity transmission method based on visual feedback, comprising the following steps: Step one, preset standard positioning coordinates: according to the ideal positioning position of the sample table in the upper cavity and the lower cavity, the first and second binocular vision systems are used to preset the standard positioning coordinates; the standard positioning coordinates include the initial positioning coordinates of the sample table in the upper cavity, the virtual safety channel coordinates and the accurate positioning coordinates in the lower cavity; Step two, system initialization: real-time monitoring of the vacuum degree of the upper cavity and the lower cavity, when the vacuum degree of the upper cavity is better than 8×10 -11 torr, and the vacuum degree of the lower cavity is better than 6×10 -11 torr, the gate valve is opened; Step three, initial positioning of the upper cavity: the first binocular vision system is used to real-time position the three-dimensional coordinates of the sample table in the upper cavity, the real-time positioning three-dimensional coordinates are compared with the pre-stored standard positioning coordinates, and the position control unit is used to drive the compound motion device for correction, so as to ensure that the sample table is at the accurate transmission starting point; Step four, safe sample transmission and path planning: after the gate valve is opened, the position control unit controls the sample table to move along the pre-stored virtual safety channel coordinates, in this process, the first binocular vision system continuously monitors to ensure that the motion trajectory deviation is within the safety tolerance; when the sample table is lowered to the preset height, the system automatically checks whether the lateral position and angle deviation are within the safety tolerance range, if the tolerance is exceeded, the system immediately starts the closed-loop control mechanism for trajectory correction, and the position satisfies the requirements before the movement can continue, so as to ensure that the transmission process is always in the preset virtual safety channel; Step five, accurate positioning of the lower cavity: when the sample table enters the lower cavity, the system enters the accurate positioning and continuous safety monitoring mode, the second binocular vision system is used to real-time three-dimensional positioning of the sample table, and the real-time three-dimensional positioning is compared with the pre-stored accurate positioning coordinates in the lower cavity, the position control unit is used to drive the compound motion device for multi-degree-of-freedom precise adjustment, until the deviation is less than the accurate positioning threshold; Step six, continuous anti-collision monitoring: the second binocular vision system continuously runs at a set frequency, and real-time calculates the spatial distance between the sample table and the analyzer; the distance data is fed back to the processing module in real time to ensure that it is always maintained above the safe working distance; when the distance is less than the first safety threshold, the system enables the feedforward control strategy to actively adjust the motion trajectory of the sample table to avoid potential collision risk; when the distance is further reduced to the second dangerous threshold, the system immediately stops all movements and triggers the sound and light alarm to ensure the safety of the equipment; Step seven, high-precision scanning and thermal compensation: after completing single-point fine positioning, the system moves the sample stage according to the preset scanning sequence, and repeats the fine positioning process in step five at each point to ensure the coordinate accuracy of each data point. The system periodically runs the positioning process to automatically compensate for the pose drift caused by thermal deformation.

[0011] Further, the method for presetting standard positioning coordinates and real-time positioning coordinates, the image acquisition unit acquires sample stage images through the first and second binocular vision systems, the information processing unit pre-processes the images, and converts the image data into three-dimensional coordinates of the sample stage based on the binocular vision principle.

[0012] Further, the pre-processing operation performed by the information processing unit on the image includes noise suppression, contrast enhancement and lens distortion correction to generate clear and standardized image frames for subsequent coordinate calculation.

[0013] Further, the method for converting image data into three-dimensional coordinates of the sample stage by the first and second binocular vision systems, the output feature points are two-dimensional pixel coordinates (u1, v1) and (u2, v2) in the camera image coordinate system; based on the pre-calibrated binocular camera parameters, the three-dimensional space coordinates (X, Y, Z) of the feature points in the world coordinate system are reconstructed by the triangulation method, and the mathematical model is defined by the following equation set: ; Where λ1 and λ2 are scale factors, M1 and M2 are camera intrinsic matrices, I represents the unit matrix / identity matrix of the image coordinate system (or image plane), and R and T are the rotation matrix and translation vector of the right camera relative to the left camera, respectively.

[0014] Further, the processing module compares the real-time positioning three-dimensional coordinates with the pre-stored standard positioning coordinates, outputs the displacement deviation instruction, and obtains the displacement deviation instruction by displacement deviation calculation. The real-time positioning three-dimensional coordinates (X, Y, Z) obtained by calculation are compared with the pre-stored standard positioning coordinates (X0, Y0, Z0), the linear displacement deviation instructions (ΔX, ΔY, ΔZ) of each axis are generated by difference calculation, and are output to the position control unit. The position control unit receives the displacement deviation instruction, calculates the driving signal of each motor in the compound motion device through the built-in motion control card, so as to realize the automatic precise correction of the position of the sample stage.

[0015] The present application has the following advantages: (1) Micron-level positioning accuracy: Through the synergistic effect of high-resolution binocular vision system and precise closed-loop control algorithm, the spatial pose of the sample stage can be accurately perceived in real time, and automatic compensation correction can be performed according to the deviation. For example, the positioning accuracy in x / y / z directions can be better than ±1 μm, and the directional positioning accuracy can be better than ±1 μm (depending on the specific design), which is several times or even an order of magnitude higher than the accuracy of traditional mechanical positioning methods (usually on the order of ±10 μm).

[0016] (2) Vacuum-compatible monitoring: Non-contact machine vision technology is used for pose measurement, and all vision components can be installed outside the vacuum chamber (observed through the vacuum window), ensuring that the vacuum environment is not disturbed.

[0017] (3) Intelligent safety guarantee: The system can monitor the relative distance between the sample stage and the surrounding key components (such as the detector) in real time, and has the ability to actively warn and avoid collisions, fundamentally eliminating the risk of equipment collision caused by human error, and significantly improving the safety of experiments.

[0018] (4) High automation and efficiency improvement: The sample transfer and positioning process realizes an automatic closed loop of "detection-decision-execution", and the user only needs to set the target pose, and the system can automatically, quickly and accurately complete the operation, greatly liberating manpower and improving experimental efficiency.

[0019] (5) Real-time compensation for thermal deformation: Continuous visual monitoring can capture and compensate for position drift caused by temperature changes in real time, ensuring the stability of the sample during long-term and variable temperature measurement, and ensuring the comparability and effectiveness of the data.

[0020] (6) Good applicability and expandability: The system design does not depend on the specific structure and form of the sample stage, and the core vision feedback and control strategy is universal. Through the adaptation of markers and adjustment of vision parameters, it can be widely deployed in different types of synchrotron radiation experimental stations or other scientific research equipment that requires precise positioning in a vacuum environment. The vision system has good adaptability to sample size and features within a certain range. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used for the embodiment description.

[0022] Figure 1 : The sample stage of the present application is located in the upper cavity structure schematic diagram.

[0023] Figure 2 : The sample stage of the present application extends to the lower cavity structure schematic diagram.

[0024] In the drawings, the components represented by each reference numeral are listed as follows: upper cavity 8, lower cavity 9, gate valve 15, polar angle rotating device 14, analyzer 10, sample rod 5, sample 6, sample table 7, X-axis horizontal displacement table 11, Y-axis horizontal displacement table 12, Z-axis vertical displacement table 13, first camera 1, second camera 2, third camera 3, and fourth camera 4. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application.

[0026] As shown in Figure 1 and Figure 2 : an automatic positioning analysis cavity of a sample table based on visual feedback, the analysis cavity has an upper cavity 8 and a lower cavity 9 inside; the lower end of the upper cavity 8 is connected to the lower cavity 9 through a first communication cavity, the first communication cavity is provided with a gate valve 15, and the upper end is installed with a composite motion device through a second communication cavity; the side wall of the lower cavity 9 is connected to an analyzer 10; The output end of the composite motion device is connected to the sample table 7 through a sample rod 5, which is used to control the sample table 7 to adjust in multiple degrees of freedom; the composite motion device includes an X-axis horizontal displacement table 11, a Y-axis horizontal displacement table 12, a Z-axis vertical displacement table 13, and a polar angle rotating device 14 located in the second communication cavity, and each is independently provided with a stepping motor for driving, which is used for three-dimensional coordinate correction to ensure that the sample is always in the position of the rotation center during rotation.

[0027] The upper cavity 8 and the lower cavity 9 are respectively provided with first and second binocular vision systems corresponding to optical windows; the first binocular vision system includes a first camera 1 and a second camera 2, which are used to synchronously collect position information of the sample table 7 in the upper cavity 8 through the first camera 1 and the second camera 2; the second binocular vision system includes a third camera 3 and a fourth camera 4, which are used to jointly solve the position information of the sample table 7 in the lower cavity 9 and the three-dimensional spatial position information relative to the analyzer 10.

[0028] The processing module includes a single-chip microcomputer or a microcomputer processor as a processing unit, has an integrated memory and an arithmetic unit; an image acquisition unit, an information processing unit, and a position control unit are integrated; The image acquisition unit is used to drive the first and second binocular vision systems to collect real-time images of visual marks of the sample table 7; The information processing unit is used to pre-process the images, and convert image data into three-dimensional coordinates and poses of the sample table based on the binocular stereo vision principle; The position control unit is used to adjust the sample table in multiple degrees of freedom through the composite motion device.

[0029] A sample stage automatic positioning analysis cavity transmission sample method based on visual feedback, comprising the following steps: Step one, preset standard positioning coordinates: according to the ideal positioning position of the sample stage 7 in the upper cavity 8 and the lower cavity 9, the first and second binocular vision systems are used to preset the standard positioning coordinates; The standard positioning coordinates include the initial positioning coordinates of the upper cavity sample stage 7, the virtual safety channel coordinates and the precise positioning coordinates of the lower cavity 9; Step two, system initialization: real-time monitoring of the vacuum degree of the upper cavity 8 and the lower cavity 9, when the vacuum degree of the upper cavity 8 is better than 8×10 -11 torr, the vacuum degree of the lower cavity 9 is better than 6×10 -11 torr, open the shutter valve 15; This double vacuum guarantee mechanism ensures that the sample will not be contaminated by gas molecules during transmission, and provides the necessary conditions for precise measurement in ultra-high vacuum environment.

[0030] Step three, initial positioning of the upper cavity: the first binocular vision system is used to real-time positioning of the sample stage 7 in the upper cavity 8 three-dimensional coordinates, the real-time positioning three-dimensional coordinates are compared with the pre-stored standard positioning coordinates, and the position control unit is used to drive the composite motion device for correction, to ensure that the sample stage is at the accurate transmission starting point; Step four, safe sample transmission and path planning: after the shutter valve 15 is opened, the position control unit controls the sample stage 7 to move along the pre-stored virtual safety channel coordinates, in this process, the first binocular vision system continues to monitor, to ensure that the motion trajectory deviation is within the safety tolerance; When the sample stage 7 is lowered to the preset height, the system automatically checks whether the horizontal position and angle deviation are within the safety tolerance range (position deviation ±2 mm, angle deviation ±0.5°), if the tolerance is exceeded, the system immediately starts the closed-loop control mechanism to correct the trajectory, and after the pose meets the requirements, the movement can continue, to ensure that the sample transmission process is always in the preset virtual safety channel; Step five, precise positioning of the lower cavity: when the sample stage 7 enters the lower cavity 9, the system enters the precise positioning and continuous safety monitoring mode, the second binocular vision system is used to real-time three-dimensional calibration of the sample stage 7, and compared with the pre-stored precise positioning coordinates of the lower cavity, the position control unit drives the composite motion device to make multi-degree-of-freedom precise adjustment, until the deviation is less than the precise positioning threshold (translation error threshold >3μm); Step six, continuous anti-collision monitoring: the second binocular vision system runs continuously at a set frequency (frequency of 30 Hz), and the spatial distance between the sample stage 7 and the analyzer 10 is calculated in real time; the distance data is fed back to the processing module in real time to ensure that it is always maintained above the safe operating distance; when the distance is less than the first safety threshold (threshold of 30 mm), the system activates the feedforward control strategy to actively adjust the sample stage motion trajectory to avoid potential collision risks; when the distance is further reduced to the second danger threshold (threshold of 15 mm), the system immediately suspends all motion and triggers an audible and visual alarm to ensure device safety; Step seven, high-precision scanning and thermal compensation: after completing single-point precision positioning, the system moves the sample stage according to the preset scanning sequence, and repeats the precision positioning process in step five at each point to ensure the coordinate accuracy of each data point. The system periodically runs the positioning process to automatically compensate for pose drift caused by thermal deformation.

[0031] Wherein, the method of preset standard positioning coordinates and real-time positioning coordinates, the image acquisition unit acquires the image of the sample stage 7 through the first and second binocular vision systems, the information processing unit pre-processes the image, and converts the image data into three-dimensional coordinates of the sample stage based on the binocular vision principle.

[0032] Wherein, the pre-processing operation performed by the information processing unit on the image includes noise suppression, contrast enhancement and lens distortion correction, to generate clear and standardized image frames for subsequent coordinate calculation.

[0033] Wherein, the method of converting image data into three-dimensional coordinates of the sample stage by the first and second binocular vision systems, the output feature points are two-dimensional pixel coordinates (u1, v1) and (u2, v2) in the camera image coordinate system; based on the pre-calibrated binocular camera parameters, the three-dimensional space coordinates (X, Y, Z) of the feature points in the world coordinate system are reconstructed by the triangulation method, and the mathematical model is defined by the following equation set: ; Wherein, λ1, λ2 are scale factors, M1, M2 are camera intrinsic matrices, I represents the unit matrix / identity matrix of the image coordinate system (or image plane), R and T are the rotation matrix and translation vector of the right camera relative to the left camera, respectively.

[0034] The processing module compares the real-time positioning three-dimensional coordinates with the pre-stored standard positioning coordinates, outputs a displacement deviation instruction, and obtains the displacement deviation instruction through displacement deviation calculation. The real-time positioning three-dimensional coordinates (X, Y, Z) obtained by solving are compared with the pre-stored standard positioning coordinates (X0, Y0, Z0), a linear displacement deviation instruction (ΔX, ΔY, ΔZ) of each axis is generated through difference calculation, and is output to a position control unit. The position control unit receives the displacement deviation instruction, solves the instruction into a driving signal of each motor in the compound motion device through a motion control card built-in the position control unit, so as to realize automatic and precise correction of the sample table position.

[0035] The embodiments are selected and specifically described in the specification in order to better explain the principles and practical applications of the present application, so that those skilled in the art can well understand and utilize the present application.

Claims

1. A sample stage automatic positioning analysis chamber based on visual feedback, characterized in that: the analysis chamber has an upper chamber (8) and a lower chamber (9) inside; the lower end of the upper chamber (8) is connected to the lower chamber (9) through a first communication chamber, the first communication chamber is provided with a gate valve (15), and the upper end is provided with a composite motion device through a second communication chamber; the side wall of the lower chamber (9) is connected to an analyzer (10); the output end of the composite motion device is connected to a sample stage (7) through a sample rod (5) for controlling the sample stage (7) to adjust in multiple degrees of freedom; the upper chamber (8) and the lower chamber (9) are respectively provided with first and second binocular vision systems corresponding to optical windows; a processing module is included, which is integrated with an image acquisition unit, an information processing unit and a position control unit; the image acquisition unit is used to drive the first and second binocular vision systems to collect real-time images of visual markers of the sample stage (7); the information processing unit is used to pre-process the images and convert the image data into three-dimensional coordinates and poses of the sample stage based on the binocular stereo vision principle; the position control unit is used to adjust the sample stage in multiple degrees of freedom through the composite motion device.

2. The visual feedback based sample stage auto-positioning analysis chamber according to claim 1, wherein: The composite motion device includes an X-axis horizontal displacement table (11), a Y-axis horizontal displacement table (12), a Z-axis vertical displacement table (13) and a polar angle rotating device (14) in the second communication chamber, and is independently provided with a stepping motor for driving, for three-dimensional coordinate correction, to ensure that the sample is always in the position of the rotation center during rotation.

3. The visual feedback based sample stage auto-positioning analysis chamber according to claim 1, wherein: The first binocular vision system includes a first camera (1) and a second camera (2) for synchronously collecting position information of the sample stage (7) in the upper chamber (8) through the first camera (1) and the second camera (2); the second binocular vision system includes a third camera (3) and a fourth camera (4) for jointly calculating the position information of the sample stage (7) in the lower chamber (9) and the three-dimensional spatial position information relative to the analyzer (10).

4. The sample transfer method of claim 1, wherein the method further comprises: determining a position of the sample stage based on the visual feedback. The following steps are included: Step one, preset standard positioning coordinates: according to the ideal positioning position of the sample stage (7) in the upper chamber (8) and the lower chamber (9), the first and second binocular vision systems are used to preset standard positioning coordinates; the standard positioning coordinates include the initial positioning coordinates of the sample stage (7) in the upper chamber, the virtual safety channel coordinates and the precise positioning coordinates of the lower chamber (9); Step two, system initialization: real-time monitoring of the vacuum degree of the upper cavity (8) and the lower cavity (9), when the vacuum degree of the upper cavity (8) is better than 8x10 -11 torr, the vacuum degree of the lower cavity (9) is better than 6x10 -11 torr, open the shutter valve (15); Step three, initial positioning of the upper chamber: the first binocular vision system is used to real-time position the three-dimensional coordinates of the sample stage (7) in the upper chamber (8), the real-time positioning three-dimensional coordinates are compared with the pre-stored standard positioning coordinates, and the composite motion device is driven by the position control unit for correction, to ensure that the sample stage is at the accurate sample transmission starting point. Step four, safe sample transmission and path planning: after the gate valve (15) is opened, the position control unit controls the sample table (7) to move along the pre-stored virtual safe channel coordinates. In this process, the first binocular vision system continuously monitors to ensure that the motion trajectory deviation is within the safety tolerance. When the sample table (7) is lowered to the preset height, the system automatically checks whether the lateral position and angle deviation are within the safety tolerance range. If the tolerance is exceeded, the system immediately starts the closed-loop control mechanism for trajectory correction. After the pose meets the requirements, the movement can continue to ensure that the sample transmission process is always within the pre-set virtual safe channel; Step five, accurate positioning in the lower cavity: when the sample table (7) enters the lower cavity (9), the system enters the accurate positioning and continuous safety monitoring mode. The second binocular vision system performs real-time three-dimensional calibration of the sample table (7) and compares it with the pre-stored accurate positioning coordinates of the lower cavity. The position control unit drives the compound motion device to perform multi-degree-of-freedom precise adjustment until the deviation is less than the accurate positioning threshold. Step six, continuous anti-collision monitoring: the second binocular vision system continuously runs at a set frequency, and calculates the spatial distance between the sample table (7) and the analyzer (10) in real time. This distance data is fed back to the processing module in real time to ensure that it is always maintained above the safe operating distance. When the distance is less than the first safety threshold, the system activates the feedforward control strategy to actively adjust the sample table motion trajectory to avoid potential collision risks. When the distance further reduces to the second danger threshold, the system immediately suspends all movements and triggers audible and visual alarms to ensure equipment safety. Step seven, high-precision scanning and thermal compensation: after completing single-point accurate positioning, the system moves the sample table according to the pre-set scanning sequence. The accurate positioning process in step five is repeated at each point to ensure the coordinate accuracy of each data point. The system periodically runs the positioning process to automatically compensate for the pose drift caused by thermal deformation.

5. The sample transfer method of claim 4, wherein the visual feedback based automatic positioning of the sample stage to the analysis chamber comprises: The method for pre-set standard positioning coordinates and real-time positioning coordinates, the image acquisition unit obtains the image of the sample table (7) through the first and second binocular vision systems, and the information processing unit pre-processes the image. Based on the binocular vision principle, the image data is converted into the three-dimensional coordinates of the sample table. ​ 6. The sample transfer method for automatically positioning the analysis chamber on the sample stage based on visual feedback according to claim 5, characterized in that: The pre-processing operation performed by the information processing unit on the image includes noise suppression, contrast enhancement, and lens distortion correction to generate clear and standardized image frames for subsequent coordinate calculation.

7. The sample transfer method of claim 6, wherein the method further comprises: determining a position of the sample stage based on the visual feedback. The method for converting image data into three-dimensional coordinates of the sample table by the first and second binocular vision systems outputs the two-dimensional pixel coordinates (u1, v1) and (u2, v2) of the feature points in the camera image coordinate system. Based on the pre-calibrated binocular camera parameters, the three-dimensional space coordinates (X, Y, Z) of the feature points in the world coordinate system are reconstructed by the triangulation method, and the mathematical model is defined by the following equation set: ； where λ1 and λ2 are scale factors, M1 and M2 are camera intrinsic matrices, I represents the unit matrix / identity matrix of the image coordinate system, and R and T are the rotation matrix and translation vector of the right camera relative to the left camera, respectively.

8. The sample transfer method of claim 7, wherein the method further comprises: determining a position of the sample stage based on the visual feedback. The processing module compares the real-time positioning three-dimensional coordinates with the pre-stored standard positioning coordinates, outputs a displacement deviation instruction, and obtains the displacement deviation instruction through displacement deviation calculation. The real-time positioning three-dimensional coordinates (X, Y, Z) obtained by solving are compared with the pre-stored standard positioning coordinates (X0, Y0, Z0), a linear displacement deviation instruction (ΔX, ΔY, ΔZ) of each axis is generated through difference calculation, and is output to a position control unit. The position control unit receives the displacement deviation instruction, solves the instruction into a driving signal of each motor in the compound motion device through a built-in motion control card, so as to realize automatic and precise correction of the sample table position.

Images