Automatic micro transfer printing system based on computer vision alignment

Abstract

The invention discloses an automatic micro transfer printing system based on computer vision alignment, which comprises a PDMS (Polydimethylsiloxane) film seal, a motion control system and a vision alignment system, the motion control system comprises a control module, a large-stroke three-axis motion sliding table and two high-precision three-degree-of-freedom electric sliding tables, a U-shaped bracket is mounted on the three-axis motion sliding table, and the vision alignment system is mounted on the U-shaped bracket; the two three-degree-of-freedom electric sliding tables are installed on two cantilevers of the U-shaped support respectively, and the two ends of the PDMS film seal are fixed to the two three-degree-of-freedom electric sliding tables respectively. The visual alignment system is used for shooting a micro device to be transferred and transmitting an image to the control module, and the control module recognizes the position of the micro device in the image and controls the PDMS film seal to move to the position over the micro device to be transferred for transfer. The contact angle between the transfer printing device and the substrate is changed by flexibly adjusting the form of the film through the two three-degree-of-freedom electric sliding tables, so that the adhesive force between the device and the substrate interface is more convenient to control, and the device is flexibly picked up and printed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of micro transfer printing technology, and particularly relates to a micro-nano device transfer printing system using computer vision for high-precision alignment. BACKGROUND

[0002] With the rapid development of modern information industry, the application of optoelectronic devices in the fields of communication interconnection, biological medicine, quantum computing, automatic driving and the like is increasingly concerned by people. Silicon-based optoelectronic technology (PIC) taking light as a carrier has obvious advantages in bandwidth, power consumption, transmission speed and cost. In optoelectronic integration, the key lies in the silicon-on-insulator (SOI) platform, which is compatible with the complementary metal-oxide-semiconductor process (CMOS) of integrated circuits in manufacturing process, and can realize optoelectronic hybrid integration on a chip. Therefore, the SOI platform is increasingly concerned and provides a highly potential platform for large-scale optoelectronic integration. However, for traditional silicon-based semiconductors, due to its indirect bandgap semiconductor material, the light-emitting efficiency is very low, and it has always been considered unsuitable for making light source materials. Therefore, it is necessary to use III-V semiconductor materials with direct bandgap to make high-efficiency light source devices. In view of this, we can use micro transfer printing method to realize the hetero-integration of active devices on passive silicon chips. Micro transfer printing technology is a new assembly technology that can transfer and assemble materials or devices on a donor substrate to a receiving substrate, and is widely used in various devices. If it is applied to micron-level devices, it is called micro transfer printing technology, so as to realize high-precision and high-density integration of various micro devices. The transfer medium used in the transfer process generally adopts elastic material stamp (PDMS), which is very suitable for processing single thin and fragile micro material and micro device samples. Therefore, in the fields of flexible wearable electronics, flexible semiconductors and micro-nano optical device integration, this micro transfer printing technology has wide application potential.

[0003] But at this stage, it is still difficult to transfer micro devices. In existing micro transfer printing technology, transfer printing needs to go through three different stages of picking up, aligning and printing. When vacuum adsorption method is used for transfer printing, the adsorption hole will be blocked in the process of adsorption transfer due to the small size of the device itself. However, the production of smaller vacuum adsorption holes requires very high production process, which greatly increases the production cost. The applicant of the patent proposed a thin film-based transfer printing technology in 2022, with publication number CN115968187A. The technology changes the morphology of the PDMS stamp by adjusting the distance between the positioning plates, increases or reduces the adhesion of the stamp to the micro device, and realizes the picking up and releasing of the transferred device. The defect of this technology is that it relies on manual operation or simple mechanical positioning, which is difficult to meet the demand of high precision and high efficiency. At the same time, the key of micro transfer printing technology is to accurately transfer micron or nanometer devices from the donor substrate to the receiver substrate, which puts forward very high requirements for the alignment accuracy. Therefore, the algorithm driven motor combined with image recognition is introduced to realize the alignment, which promotes its application in high-end manufacturing field SUMMARY

[0004] In view of the problems of low control efficiency and insufficient positioning accuracy existing in the existing transfer printing technology, the present application proposes a high-precision alignment system using computer vision technology and algorithm-driven motor two modes.

[0005] The technical scheme adopted by the present application is: An automatic micro transfer printing system based on computer vision alignment, comprising a PDMS film stamp, a motion control system and a visual alignment system. The motion control system comprises a control module, a three-axis motion slide with large stroke and two three-degree-of-freedom electric slides. A U-shaped support is installed on the three-axis motion slide, and the two three-degree-of-freedom electric slides are respectively installed on the two cantilevers of the U-shaped support. The two ends of the PDMS film stamp are respectively fixed on the two three-degree-of-freedom electric slides. The visual alignment system is used to shoot the micro device to be transferred, and the image is transmitted to the control module. The control module identifies the position of the micro device in the image and controls the PDMS film stamp to move to the top of the micro device to be transferred for transfer printing.

[0006] The three-axis motion slide with large stroke comprises a Z-axis slide 1, a Y-axis slide 2 and an X-axis slide 3. The U-shaped support is installed on the Z-axis slide 1, and the U-shaped support is composed of a cross beam and a left displacement adjusting frame 6 and a right displacement adjusting frame 7.

[0007] The motion control system and the visual alignment system are on the optical vibration isolation platform 10. The camera support 11 is installed on the optical vibration isolation platform 10, the industrial camera 12 is installed on the camera support 11, and the microscope lens 13 is installed on the industrial camera 12. The three-axis motion slide is provided with a sample stage 14 below the U-shaped support.

[0008] The transfer process comprises: I. The device to be transferred is located on the source substrate, and the PDMS film stamp is moved to above the device to be transferred by a large-stroke three-axis motion slide table; II. Two high-precision three-degree-of-freedom electric slides are controlled to approach, so that the PDMS film stamp is bent to a preset curvature, then the two three-degree-of-freedom electric slides are synchronously lowered, the stamp is moved downward, the lowest point of the bent stamp contacts the device to be transferred, and finally the PDMS film is lifted to pick up the device to be transferred; III. The PDMS film stamp is moved to above the target substrate by a large-stroke three-axis motion slide table, a vision alignment system photographs the device to be transferred and the target substrate, the receiving positions of the device to be transferred and the target substrate are identified through image recognition, the center distance between the two is calculated, the offset amounts of the X and Y axes are obtained, and the positions of the PDMS film stamp are adjusted by controlling the two three-degree-of-freedom electric slides so that the center of the transferred device coincides with the center of the receiving position; IV. The three-degree-of-freedom electric slides at both ends of the PDMS film stamp are lowered in sequence to make the device to be transferred reach the receiving position, and then the three-degree-of-freedom electric slides at both ends of the PDMS film are lifted in sequence to release the transferred device.

[0009] The PDMS film stamp is lowered to make the device to be transferred reach the receiving position, the stamp with the transferred device is slowly lowered, and the transferred device is slowly released. The lifting speed of the PDMS film stamp when picking up the transferred device in step II is greater than the lifting speed of the PDMS film stamp in step IV.

[0010] In step III, the identification of the receiving positions of the transferred device and the target substrate comprises: (1) The gray-scale image is converted into a binary image, and the best threshold value for binarization is determined by using the OTSU algorithm; (2) For the receiving position, open operation and close operation are used in sequence to clean the image edges and fill small gaps in the image, and the image is subjected to smoothing treatment by using Gaussian filtering; (3) Clear edge information is extracted from the image by an edge detection module in the algorithm, and finally the profile is screened and drawn; (4) Center marking of the device to be transferred: the profile of the largest object in the image is detected and extracted, the center and radius of the smallest circumscribed circle are calculated and marked, the largest profile is subjected to geometric analysis, the smallest circle surrounding it is obtained, and the center position is marked; (5) Center marking of the receiving position: the largest external profile of the receiving position region is detected, and the smallest circumscribed rectangle surrounding the profile is calculated. The width-to-height ratio of the rectangle is normalized according to the standard, and the center point of the rectangle is calculated and marked.

[0011] Compared with the prior art, the present application has the following advantages: 1、The contact angle between the transfer device and the substrate can be changed by automatically and flexibly adjusting the shape of the film through two three-degree-of-freedom electric sliding tables, the adhesion between the device and the substrate interface is more convenient to control, and the pickup and printing of the device are flexibly realized.

[0012] 2、The visual alignment system is mainly responsible for real-time monitoring of the alignment process of the transfer device and the receiving position on the target substrate, so as to complete the high-precision alignment work of micro transfer. The computer analyzes the collected images, obtains the relative position deviation at this time, and controls the displacement table to adjust and compensate.

[0013] 3、Compared with the existing micro transfer equipment, while completing the basic pickup, alignment, printing and other work of micro transfer, the cost is effectively reduced, and the transfer success rate is higher, which is more convenient for actual operation. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 The device system block diagram, the system is mainly composed of motion control system, PDMS film stamp and visual alignment system three parts.

[0015] Figure 2 The proposed micro transfer device schematic diagram, 1-Z axis sliding table; 2-Y axis sliding table; 3-X axis sliding table; 4-left three-degree-of-freedom linear electric sliding table; 5-right three-degree-of-freedom linear electric sliding table 2; 6-left displacement adjustment frame; 7-right displacement adjustment frame; 8-displacement table motion control module; 9-transfer stamp; 10-optical vibration isolation platform; 11-camera support; 12-industrial camera; 13-microscope lens; 14-manual displacement stage; 15-manual adjustment shaft.

[0016] Figure 3 The schematic diagram of the micro transfer process demonstrated, Figure 3 (a) is a schematic diagram of 9-stamp contacting 16-device to be transferred on 14-source substrate, Figure 3 (b) is a schematic diagram of quickly lifting 5-displacement adjustment frame after contacting to complete the pickup action, Figure 3 (c) is a schematic diagram of 16-device to be transferred and 14-target substrate fully contacting the receiving position to perform the printing process, Figure 3 (d) is a schematic diagram of slowly lifting 9-stamp to complete the printing action. The whole process is assisted by 13-industrial camera throughout the alignment.

[0017] Figure 4 The grating and photodetector identification results.

[0018] Figure 5 The alignment results obtained by the alignment system. DETAILED DESCRIPTION

[0019] The application will be described in more detail with reference to the drawings and embodiments thereof, but the scope of protection is not limited to the described.

[0020] Embodiment 1 An automatic micro-transferring system based on computer vision alignment, comprising a PDMS film stamp, a motion control system and a vision alignment system, the motion control system comprising a control module, a large-stroke three-axis motion slide and two high-precision three-degree-of-freedom electric slides, the large-stroke three-axis motion slide comprising a Z-axis slide 1, a Y-axis slide 2 and an X-axis slide 3, a U-shaped support being mounted on the Z-axis slide 1, the U-shaped support being composed of a crossbeam and a left displacement adjusting frame 6 and a right displacement adjusting frame 7.

[0021] The two three-degree-of-freedom electric slides 4 and 5 are respectively mounted on the left displacement adjusting frame 6 and the right displacement adjusting frame 7 of the U-shaped support, and the two ends of the PDMS film stamp are respectively fixed on the two three-degree-of-freedom electric slides, and the shape of the film can be adjusted by changing the distance between the two three-degree-of-freedom electric slides.

[0022] The vision alignment system is used for shooting the micro device to be transferred, and the image is transmitted to the control module, the control module identifies the position of the micro device in the image and controls the PDMS film stamp to move to the top of the micro device to be transferred for transferring.

[0023] The motion control system and the vision alignment system are on an optical vibration isolation platform 10, a camera support 11 is mounted on the optical vibration isolation platform 10, an industrial camera 12 is mounted on the camera support 11, and a microscope lens 13 is mounted on the industrial camera 12; a stage 14 is arranged below the U-shaped support of the three-axis motion slide.

[0024] The transferring process comprises: I. The device to be transferred is located on a source substrate, and the PDMS film stamp is moved to the top of the device to be transferred by the large-stroke three-axis motion slide; II. The two high-precision three-degree-of-freedom electric slides are controlled to approach each other, so that the PDMS film stamp is bent to a preset curvature, the stamp is moved downward, the lowest point of the bent stamp contacts the device to be transferred, and then the PDMS film is lifted to complete the pickup of the device to be transferred; III. The PDMS film stamp is moved to the top of a target substrate by the large-stroke three-axis motion slide, the vision alignment system shoots the device to be transferred and the target substrate, the receiving position of the device to be transferred and the target substrate is identified through image recognition, the center distance between the two is calculated, the offset of the X-axis and the Y-axis is obtained, the position of the PDMS film is adjusted by controlling the two three-degree-of-freedom electric slides so that the center of the transferred device coincides with the center of the receiving position, and the alignment operation is completed; Four, the three degrees of freedom motorized slides at both ends of the PDMS film stamp are lowered in turn to make the device to be transferred to reach the receiving position, and then the three degrees of freedom motorized slides at both ends of the PDMS film stamp are raised in turn to release the transferred device. There is a problem of breaking of Van der Waals force in the middle, because the film is sticky and the receptor does not have stickiness, if the film is used to transfer the device, it is difficult to break the Van der Waals force by going straight up and straight down, which causes difficulty in releasing and reduces the success rate, therefore the two ends of the PDMS film stamp need to be lowered and raised in turn.

[0025] The identification of the receiving position of the transferred device and the target substrate in step three comprises: (1) converting the gray-scale image into a binary image, and determining the best threshold value of the binary image by using the OTSU algorithm; (2) for the receiving position, using open operation and close operation in turn to clean the image edge and fill the small gap in the image, and using Gaussian filtering to smooth the image; (3) extracting clear edge information from the image through the edge detection module in the algorithm, and finally screening the contour and drawing; (4) center marking of the device to be transferred: detecting and extracting the contour of the largest object in the image, calculating and marking the center and radius of the smallest circumscribed circle, and performing geometric analysis on the largest contour to obtain the smallest circle surrounding it and marking the center position; (5) center marking of the receiving position: detecting the largest external contour of the receiving position area, and calculating the smallest circumscribed rectangle surrounding the contour. Calculate the aspect ratio of the rectangle, normalize according to the standard, calculate and mark the center point of the rectangle.

[0026] Example 2: The micro transfer device provided by the application is composed of a motion control system, a transfer execution system and a visual alignment system, as shown in Figure 1 The motion control system is responsible for regulating the relative position of the transfer stamp and the substrate plane, and cooperates with the visual alignment system to adjust the relative position of the stamp and the substrate in the vertical direction through the double-arm structure on the large-stroke displacement shaft. The visual alignment system uses computer vision technology and algorithm-driven motor linkage strategy to achieve high-precision alignment. First, the image processing technology is used to process the collected image, and the motor is driven to realize static vision and algorithm-driven automatic alignment. The transfer execution system adjusts the PDMS film stamp shape flexibly and automatically by installing a high-precision displacement table on the double-arm, and completes the picking and printing of the device. Finally, the three systems work together to realize high-precision micro transfer.

[0027] In the application, as Figure 2As shown, the motion control system is composed of three large-stroke electric slides (1, 2, 3) and two high-precision three-degree-of-freedom linear electric slides (4, 5), wherein the large-stroke electric slides are used to complete the displacement from the original substrate to the target substrate, and the high-precision three-degree-of-freedom linear electric slides are used to automatically adjust the morphology of the transfer PDMS film stamp. The large-stroke electric slide splicing extension double arm with high-precision three-degree-of-freedom linear electric slide simulates the manual peeling method of both hands and completes the high-precision alignment of the transfer device and the target substrate with the visual alignment system. Finally, the device is picked up by the high-speed large-stroke electric slide, and the device is printed by the high-precision three-degree-of-freedom linear electric slide. During the whole process, the large-stroke displacement shaft drives the high-precision displacement shaft to move.

[0028] The transfer execution system is mainly responsible for realizing the picking and printing work in the process. The present application adopts a flexible PDMS stamp and uses the structure in the previous patent (Invention name: Thin film-based transfer printing technology; Patent number ZL 202211564916.0). Compared with the rigid stamp used in the past, the transfer process is more flexible. The morphology of the film can be automatically and flexibly adjusted by the transfer device mentioned in this research to change the contact angle between the transfer device and the substrate, which is more convenient for controlling the adhesion between the device and the substrate interface, and flexibly realizes the picking and printing of the device.

[0029] The visual alignment system is mainly responsible for real-time monitoring of the alignment process of the transfer device and the receiving position on the target substrate, in order to complete the high-precision alignment work of micro transfer. The visual alignment system is mainly composed of an industrial camera 12 and a microscope lens 13, which are fixed on an optical vibration isolation platform 10 by a camera support 11 for adjusting their relative positions. The industrial camera is used to collect the images of the transfer device and the receiving substrate during the transfer process, and the collected images are analyzed by the computer to obtain the relative position deviation at this time, and the displacement table is controlled to adjust and compensate.

[0030] Figure 2 The structure diagram of the transfer device proposed by the present application, which can be used for picking and transferring the device to be transferred. The device to be transferred can be a resistor or other optoelectronic chip element. In this embodiment, it is a detector.

[0031] First, the device to be transferred is located on the source substrate, such as Figure 3(a) is shown, by controlling the three large stroke electric slide adjustment displacement adjustment frame 7 to move the transfer stamp 9 to the top of the device to be transferred, and then by adjusting the double-arm high-precision three-degree-of-freedom linear electric slide composed of groups 4 and 5, the stamp forms a "U" type bending, and then by adjusting the high-precision three-degree-of-freedom linear electric slide, the patch component is aligned with the lowest point of the "U" type bending of the stamp, and the stamp is lowered to contact the device to be transferred on the source substrate. After the stamp is in sufficient contact with the device to be transferred, the three large stroke electric slides are quickly raised to drive the double-arm to complete the picking action, as shown in Figure 3 (b) is shown. Next, by adjusting the high-precision three-degree-of-freedom linear electric slide, the transfer stamp with the picked device to be transferred is moved to the above of the receiving position of the target substrate 14, a coarse alignment is performed, then the double-arm high-precision three-degree-of-freedom linear electric slide is adjusted to complete the fine alignment of the device to be transferred with the receiving position. After the alignment is completed, the high-precision three-degree-of-freedom linear electric slide is continuously adjusted to slowly lower the stamp with the device to be transferred to approach the receiving position, the left three-degree-of-freedom linear electric slide 4 is lowered first to let the device to be transferred contact the receiving position of the target substrate, then the right three-degree-of-freedom linear electric slide 5 is lowered to let the device to be transferred fully contact the receiving position of the target substrate. As shown in Figure 3 (c) is shown. Finally, the two high-precision three-degree-of-freedom linear electric slides are adjusted in sequence to be raised, so that the left and right ends of the stamp are slowly moved up, and the printing of the device can be completed, as shown in Figure 3 (d) is shown.

[0032] The visual alignment system mainly consists of an industrial camera 12 and a microscope lens 13, which are fixed on an optical vibration isolation platform 10 by a camera support 11 for adjusting their relative positions. In the alignment experiment, we use computer assistance to complete high-precision alignment, that is, we use algorithms to drive the two high-precision three-degree-of-freedom linear electric slides (4, 5) to realize picking and releasing actions. When picking, the left and right slides are raised and lowered synchronously, and when releasing, the action is left down-right down-left up-right up. At the same time, image matching alignment is completed by combining static recognition in vision. In this study, the alignment system can be completed in two parts.

[0033] First, the image recognition and processing process. In the image processing algorithm, the characteristics of the equipment shooting image should be fully considered, such as resolution, noise level, etc., at the same time, the steps of image preprocessing, center position recognition, etc. are completed. The specific steps are as follows: Step 1: Use an industrial camera to collect the images of the transfer device (photodetector) and the receiving substrate (chip) in the transfer process; Step 2: Read the image and set the target area (ROI); Step 3: Process the acquired image: 1) Binaryzation of the image. That is, converting the gray image to a binary image, and automatically determining the best threshold for binaryzation using OTSU method. By this, the complexity of the image is reduced, making the subsequent edge detection easier and more efficient. If the contrast of the binaryzation image is still insufficient, the contrast can be further enhanced; 2) Morphological operation (only used when processing the grating part): first, using open operation and close operation to clean the image edge and fill the small gap in the image, second, using Gaussian filter to smooth the image and reduce the noise in the image; Step 4: Screening the contour and drawing by improving the canny edge detection algorithm, that is, extracting the clear edge information from the image by the edge detection module in the algorithm, finally screening the contour and drawing, which can be seen from Figure 4 ; Step 5: Marking and positioning the center position by the edge detection result: 1) Detector center marking: detecting and extracting the contour of the largest object in the image, calculating and marking the center and radius of the minimum circumscribed circle, and performing geometric analysis on the largest contour to obtain the minimum circle surrounding it and marking the center position; 2) Grating center marking: detecting the largest external contour of the grating area and calculating the minimum circumscribed rectangle surrounding the contour. Calculate the aspect ratio of the rectangle, normalize it according to the standard, and calculate and mark the center point of the rectangle; Step 6: Calculate the horizontal and vertical coordinates required to align the detector and grating center points and output.

[0034] Secondly, according to the results of the first part of image processing, the process of driving the motor by algorithm. First, complete the motor control function encapsulation and dynamic link library management, the main task of this part is to coordinate the motor control function, and realize the static loading and release of dynamic link library (DLL) through explicit calling. Specifically, the functions in the DLL need to be encapsulated to use these functions efficiently and safely in the program. After encapsulation, the moving distance obtained by the image processing module is used as a parameter to call the encapsulated function, so as to realize the initialization and accurate movement of the motor. The motor control steps are as follows: Step 1: Initialize parameters and control card; Step 2: Set control parameters, such as axis number and moving distance (unit: microns); Step 3: Call the Windows API function LoadLibraryW() to dynamically load the DLL; Step 4: Call the Windows API function GetProcAddress() to get the function pointer of the function to be called in the DLL, and use the function pointer to call the function in the DLL to complete the corresponding function; Step 5 At the end of the program or when no longer using functions in the DLL, call the Windows API function FreeLibrary() to release the dynamic link library. Step 6 Encapsulate the motor function Motor_move, which encapsulates the initialization function of the motor control axis and control signal, and the motion function of controlling the moving speed, direction, and distance. After encapsulation, it is called every time the movement is performed. Step 7 Call the motion instruction and query the motion state.

[0035] In the design phase of the motion function, the primary task is to build an accurate speed control function, which contains three key parameters: medium speed, acceleration, and deceleration, all measured in pulse counts. It is important to note that there may be errors between the theoretical speed setting value and the actual pulse frequency output by the motion controller. This error is due to two control variables of the pulse generation system: pulse resolution (Pulse Resolution) and multiplier (Multiplier). Specifically, an increase in the output multiplier will result in a positive correlation between the system error and the increase.

[0036] At the same time, two typical motion modes are considered: constant velocity mode (Constant Velocity Mode) and trapezoidal velocity profile mode (Trapezoidal Velocity Profile Mode). After completing the modeling of the basic motion function, the origin regression function based on position feedback is further designed to achieve precise positioning control of the system.

[0037] After completing the code integration, we use the existing program to perform multi-dimensional basic testing on the device to verify the effectiveness and robustness of the program. First, we verify the fixed distance movement function: horizontal and vertical movement tests of different ranges are performed, and through algorithm debugging, the device can accurately reach the target position. Subsequently, we test the scaling movement function, such as scaling the movement distance by 1:2 or 5:1, and the deviation between the actual displacement and the theoretical value is less than 1.5% of the set value. In addition, we also verify the multi-axis cooperative work ability, confirming that the horizontal and vertical movement sequence is consistent with the code instruction, and there is no phenomenon of axis action misplacement.

[0038] We test the alignment system in actual experiments, and the expected alignment experiment results are detailed in Figure 5 , obviously the center of the photodetector is basically coincident with the center of the target grating. In order to accurately measure the relative displacement error between the grating and the detector, we perform testing and analysis. Figure 5 The initial positions of the grating and the detector are shown, at which time the two are in an unaligned state. At the same time, Figure 5The physical distance of the required movement after the two target pattern recognition is shown, and as the experiment proceeds, the grating moves and finally aligns the result as shown. Finally, by calculation, we get the physical distance corresponding to the experimental error of the system is about 1 micron. This result shows that our method can effectively complete the micro transfer alignment operation and reduce the error, which provides a reliable foundation for subsequent high-precision alignment.

[0039] After the alignment is completed, the PDMS flexible stamp is controlled by the automatic control equipment of the transfer execution system to complete the micro transfer process. Finally, through the cooperation of the three systems, high-precision micro transfer integration of optoelectronic devices is realized.

Claims

1. An automated microtransfer printing system based on computer vision alignment, comprising a PDMS thin film stamp, a motion control system, and a vision alignment system, characterized in that: The motion control system comprises a control module, a large-stroke three-axis motion slide table and two high-precision three-degree-of-freedom electric slide tables, a U-shaped support is installed on the three-axis motion slide table, the two three-degree-of-freedom electric slide tables are respectively installed on two cantilevers of the U-shaped support, and the two ends of the PDMS film stamp are respectively fixed on the two three-degree-of-freedom electric slide tables; the visual alignment system is used for shooting a micro device to be transferred, and the image is transmitted to the control module, the control module identifies the position of the micro device in the image and controls the PDMS film stamp to move to the top of the micro device to be transferred for transfer.

2. The automated microtransfer printing system based on computer vision alignment of claim 1, wherein: The large-stroke three-axis motion slide table comprises a Z-axis slide table (1), a Y-axis slide table (2) and an X-axis slide table (3), and the U-shaped support is installed on the Z-axis slide table (1), and the U-shaped support is composed of a cross beam and a left displacement adjusting frame (6) and a right displacement adjusting frame (7).

3. The automated microtransfer printing system based on computer vision alignment of claim 2, wherein: The motion control system and the visual alignment system are on an optical vibration isolation platform (10), a camera support (11) is installed on the optical vibration isolation platform (10), an industrial camera (12) is installed on the camera support (11), and a microscope lens (13) is installed on the industrial camera (12); and a stage (14) is arranged below the U-shaped support of the three-axis motion slide table.

4. The automated microtransfer printing system based on computer vision alignment of claim 1, wherein The transfer process comprises the following steps: (1) the device to be transferred is located on a source substrate, the PDMS film stamp is moved to the top of the device to be transferred by the large-stroke three-axis motion slide table; (2) the two high-precision three-degree-of-freedom electric slide tables are controlled to approach each other, the PDMS film stamp is bent to a preset curvature, then the two three-degree-of-freedom electric slide tables are synchronously lowered, the stamp is moved downward, the lowest point of the bent stamp contacts the device to be transferred, and finally the PDMS film stamp is lifted to complete the pickup of the device to be transferred; (3) the PDMS film stamp is moved to the top of a target substrate by the large-stroke three-axis motion slide table, the visual alignment system shoots the device to be transferred and the target substrate, the receiving positions of the device to be transferred and the target substrate are identified through image recognition, the center distance between the two is calculated, the offset amounts of the X-axis and the Y-axis are obtained, the positions of the PDMS film are adjusted by controlling the two three-degree-of-freedom electric slide tables, so that the center of the transferred device coincides with the center of the receiving position; (4) the three-degree-of-freedom electric slide tables at the two ends of the PDMS film are controlled to make the device to be transferred reach the receiving position in sequence, and then the three-degree-of-freedom electric slide tables at the two ends of the PDMS film are lifted in sequence to release the device to be transferred.

5. The automated microtransfer printing system based on computer vision alignment of claim 4, wherein: The lifting speed of the PDMS film stamp in step (2) is greater than the lifting speed of the PDMS film stamp in step (4).

6. The automated microtransfer printing system based on computer vision alignment of claim 4, wherein The identification of the receiving positions of the transferred device and the target substrate in step (3) comprises the following steps: (1) the gray-scale image is converted into a binary image, and the OTSU algorithm is used to determine the best threshold value of the binary image; (2) for the receiving position, open operation and close operation are used in sequence to clean the image edges and fill the small gaps in the image, and the image is smoothed by using a Gaussian filter; (3) clear edge information is extracted from the image by using an edge detection module in the algorithm, and finally the profile is screened and drawn. (4) Center mark of the device to be transferred: Detect and extract the contour of the largest object in the image, calculate and mark the center and radius of the minimum circumscribed circle, and perform geometric analysis on the largest contour to obtain the smallest circle surrounding it and mark the center position; (5) Center mark of the receiving position: Detect the largest external contour of the receiving position area, and calculate the minimum circumscribed rectangle surrounding the contour. Calculate the aspect ratio of the rectangle, normalize it according to the standard, and calculate and mark the center point of the rectangle.

7. The automated microtransfer printing system based on computer vision alignment of claim 6, wherein: Before performing step four, take a photo again to confirm whether the center of the device to be transferred coincides with the center of the receiving position.

Images