Cell recognition method and system based on visible light imaging and Raman spectrum detection

Abstract

The invention provides a cell identification method and system based on visible light imaging and Raman spectrum detection, and the method comprises the following steps: placing a target cell sample on an object placing table, and installing a probe and a capillary microtube; acquiring a real-time display image of the target cell sample, performing target detection on the real-time display image based on the improved YOLO11 model, and determining a target cell detection result; the upper computer is connected with the displacement platform interface, the object placing table is moved, focusing and dimming operations are carried out, Raman spectrum data of target cells are obtained, the Raman spectrum data are processed and analyzed to obtain a cell recognition result, and the cell recognition result is corrected based on a YOLO11 model to obtain a final cell recognition result; setting linear actuator communication ports in sequence, and performing conversion between a screen pane coordinate system and a displacement platform plane rectangular coordinate system; according to the Raman spectrum data and the final cell recognition result, operation path planning of the mechanical arm operation module is conducted, and the optimal operation path is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of cell detection, and particularly to a cell recognition method and system based on visible light imaging and Raman spectroscopy detection. Background Art

[0002] In the field of cell detection, traditional cell detection methods mainly rely on manual operation and some basic instrument devices. For example, cell observation and counting are carried out under a microscope. This method is not only inefficient, but also requires high professional skills of operators, and is easily affected by human factors, resulting in poor accuracy and repeatability of detection results. With the development of technology, some automated cell detection technologies have gradually emerged, such as cell detection systems based on image recognition, which use computer vision algorithms to analyze and recognize cell images. However, these technologies still have many problems when dealing with complex samples.

[0003] When facing samples with rough surfaces or impurities, traditional cell detection technologies often have difficulty accurately identifying cells. The surface roughness and impurities will interfere with the acquisition and analysis of images, resulting in the masking or misjudgment of cell features. For example, in the field of criminal investigation, cell samples on items such as clothes and murder weapons extracted from the crime scene are usually contaminated by impurities such as dust and fibers, making the detection and analysis of cells extremely difficult. Moreover, existing cell detection technologies often lack sufficient precision and automation when performing further operations on cells, such as puncture, transfer, and injection, and are likely to damage cells, affecting subsequent detection and research results.

[0004] As a non-destructive detection technology, Raman spectroscopy technology can provide chemical composition and structural information of cells and has important application value in cell detection. However, the current Raman spectroscopy detection process often requires manual operation, including sample positioning, focusing, and light adjustment, etc. This is not only inefficient, but also easily affected by human factors, resulting in the accuracy and reliability of detection results being affected. In addition, the data processing and analysis of Raman spectroscopy detection are relatively complex and require professional knowledge and skills, which limits its popularization and application in practical applications. Summary of the Invention

[0005] The present invention provides a cell recognition method and system based on visible light imaging and Raman spectroscopy detection to solve the defects existing in the prior art. Through the deep cooperation of hardware-software-algorithm, the full-process automation from cell recognition to function verification is realized. Visual control of multiple robotic arms is used to perform operations such as cell transfer, puncture, and injection, and a vacuum and temperature-controllable measurement environment is provided, avoiding many risks existing in operating and transferring samples in the atmospheric environment.

[0006] In a first aspect, the present invention provides a cell recognition method based on visible light imaging and Raman spectroscopy detection, including: Obtain a target cell sample, place the target cell sample on a placement table, and install the probe and capillary microtube of the robotic arm; Connect the host computer to the microscope camera, obtain a real-time display image of the target cell sample, perform target detection on the real-time display image based on the improved YOLO11 model, and determine the target cell detection result; Connect the host computer to the displacement platform interface, move the placement table and perform focusing and lighting adjustment operations, obtain the Raman spectroscopy data of the target cells, process and analyze the Raman spectroscopy data to obtain the cell recognition result, and correct the cell recognition result based on the YOLO11 model to obtain the final cell recognition result; Set the linear actuator communication port in sequence, and perform the conversion between the screen pane coordinate system and the displacement platform plane rectangular coordinate system; According to the Raman spectroscopy data and the final cell recognition result, perform the operation path planning of the robotic arm operation module to obtain the optimal operation path to achieve the preset operation on the target cells.

[0007] According to the cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention, connecting the host computer to the microscope camera and obtaining the real-time display image of the target cell sample includes: Install python on the host computer to achieve the remote control operation of the placement table, and control the placement table to translate in the preset moving direction; After the placement table completes translation, the microscope camera automatically focuses and collects the real-time display image, and uploads it to the host computer.

[0008] According to the cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention, performing target detection on the real-time display image based on the improved YOLO11 model and determining the target cell detection result includes: Add the CBAM attention mechanism to the YOLO11 model to construct an improved YOLO11 model; The CBAM attention mechanism includes a channel attention module and a spatial attention module; Adopt the improved YOLO11 model to screen the detection frames below the preset confidence level, mark the remaining detection frames as target detection frames, obtain the position information of the target detection frames, and obtain the target cell detection result.

[0009] According to the cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention, the channel attention module includes the following steps: Perform global average pooling and global maximum pooling operations on the input feature map; Use a fully connected layer multi-layer perceptron MLP to learn the weights of each channel; Apply the learned weights to the input feature map to obtain the feature map after channel attention adjustment; Correspondingly, the spatial attention module includes the following steps: Perform global average pooling and global maximum pooling operations on the input feature map; Add the pooled feature maps by channel to obtain two 1D vectors; Perform a dot product on the two 1D vectors to form an attention weight matrix; Apply the attention weight matrix to the input feature map to obtain the feature map after spatial attention adjustment.

[0010] According to a cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention, sequentially set the communication ports of the linear actuator, and perform the conversion between the screen window coordinate system and the displacement platform plane rectangular coordinate system, including: Control the probe to perform a preset displacement on the x-axis, and obtain the pre-movement actuator coordinates and post-movement actuator coordinates of the linear actuator corresponding before and after the displacement; Obtain the pre-movement probe coordinates and post-movement probe coordinates of the probe tip on the screen window before and after the displacement; From the pre-movement probe coordinates, post-movement probe coordinates and preset displacement, obtain the mapping relationship formula between the screen window coordinate system and the displacement platform plane rectangular coordinate system; Based on the pre-movement probe coordinates and post-movement probe coordinates, determine the rotation angle of the coordinates of the probe tip in the screen window coordinate system; From the sine and cosine values of the rotation angle, as well as the pre-movement probe coordinates, post-movement probe coordinates, pre-movement actuator coordinates and the mapping relationship formula, calculate the displacement distance of the linear actuator in the displacement platform plane rectangular coordinate system.

[0011] According to a cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention, the preset operations include puncture, transfer and injection.

[0012] In a second aspect, the present invention also provides a cell recognition system based on visible light imaging and Raman spectroscopy detection, including: A cell imaging and Raman spectroscopy detection module, an automated operation module, a machine learning algorithm module and a user interaction interface; The cell imaging and Raman spectroscopy detection module includes a confocal Raman microscope and a cell recognition system; The automated operation module includes a host computer and a user operation interaction program, a displacement platform, a linear actuator and its controller, a capillary microtube, a probe and a fixture; The machine learning algorithm module includes detecting target cells using an improved YOLO11 model; The user interface is used for the user to control the system and display the results.

[0013] According to a cell recognition system based on visible light imaging and Raman spectroscopy detection provided by the present invention, the fixture is connected to the three-dimensional displacement platform through a metal cantilever to form a manipulator. A linear actuator is used to replace the manual operating rod of the three-dimensional displacement platform, and the linear motion output by the actuator is transmitted to the three-dimensional displacement platform to achieve nanoscale operation.

[0014] According to a cell recognition system based on visible light imaging and Raman spectroscopy detection provided by the present invention, the cell recognition system is used to automatically control the movement, focusing, and dimming operations of the displacement platform.

[0015] According to a cell recognition system based on visible light imaging and Raman spectroscopy detection provided by the present invention, the machine learning algorithm module is used to preprocess the Raman spectroscopy data output by the cell recognition system, and use machine learning algorithms to analyze the processed spectroscopy data to identify the type and state of cells.

[0016] The cell recognition method and system based on visible light imaging and Raman spectroscopy detection provided by the present invention, through the adoption of full automation control, the user only needs to operate in the human-machine interface carried on the upper computer to complete operations such as cell puncture, transfer, and injection, greatly reducing the difficulty of manual operation and solving the problem of time-consuming and laborious manual cell operation; and adopting multi-modal data fusion to improve the recognition accuracy. Traditional methods mostly rely on a single data source to identify cells, which is prone to misjudgment. The present invention innovatively integrates cell morphological features and Raman spectroscopy metabolic features, and uses an improved machine learning algorithm to be able to more accurately identify cell types and functional states; it can automatically identify large-area samples, solving the problem of low recognition accuracy when the sample background is complex and cells overlap and cover each other. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 is one of the flow schematic diagrams of the cell recognition method based on visible light imaging and Raman spectroscopy detection provided by the present invention; Figure 2It is the second flow schematic diagram of the cell recognition method based on visible light imaging and Raman spectroscopy provided by the present invention; Figure 3 It is the schematic diagram of the CBAM attention mechanism provided by the present invention; Figure 4 It is the schematic diagram of the robotic arm provided by the present invention; Figure 5 It is the schematic diagram of the probe and fixture provided by the present invention; Figure 6 It is the schematic diagram of the displacement platform provided by the present invention; Figure 7 It is the schematic diagram of the capillary microtube provided by the present invention. Detailed implementation manners

[0019] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] The cell recognition method and system based on visible light imaging and Raman spectroscopy provided by the present invention are applicable to single cell recognition, function verification and automated processing, and can be applied to scenarios such as disease diagnosis, microbial resource screening or drug research and development.

[0021] Figure 1 It is one of the flow schematic diagrams of the cell recognition method based on visible light imaging and Raman spectroscopy provided by the embodiments of the present invention, as Figure 1 shown, including: Step 100: Obtain a target cell sample, place the target cell sample on a placement table, and install the probe and capillary microtube of the robotic arm; Step 200: Connect the host computer to the microscopic camera, obtain a real-time display image of the target cell sample, perform target detection on the real-time display image based on the improved YOLO11 model, and determine the target cell detection result; Step 300: Connect the host computer to the displacement platform interface, move the placement table and perform focusing and light adjustment operations, obtain the Raman spectral data of the target cell, process and analyze the Raman spectral data to obtain a cell recognition result, and correct the cell recognition result based on the YOLO11 model to obtain the final cell recognition result; Step 400: Set the linear actuator communication port in sequence, and perform the conversion between the screen pane coordinate system and the displacement platform plane rectangular coordinate system; Step 500: According to the Raman spectroscopy data and the final cell recognition result, perform the operation path planning of the robotic arm operation module to obtain the optimal operation path, so as to achieve the preset operation on the target cell.

[0022] Specifically, as Figure 2 shown in the more refined flowchart, including: Step 1: Place the prepared cell sample on the stage of the Raman microscope. Install the probe and capillary microtube of the robotic arm.

[0023] Step 2: To achieve high-throughput and automated cell detection, the host computer needs to be connected to the microscopic camera to obtain real-time display images. Install python on the host computer to achieve remote operation of the Raman microscope stage. The control platform moves horizontally from left to right and vertically from top to bottom. After each translation, the microscopic camera automatically focuses and takes pictures and transmits them to the host computer. First, splice all the taken pictures, and then use the improved YOLO11 model for target detection, and filter out the detection frames with a confidence level below 0.5. Mark the remaining detection frames as suspected cells and record the position information of the detection frames.

[0024] Among them, the improved YOLO11 model introduces the CBAM attention mechanism in feature extraction, enhances the attention and extraction ability of cell features, and optimizes the network structure. The data augmentation technology is used to improve the generalization ability of the model. The CBAM structure is as Figure 3 shown. CBAM consists of a channel attention module (Channel Attention Module, CAM) and a spatial attention module (Spatial Attention Module, SAM). The design concept of CBAM is to let the model first focus on the importance at the channel level, perform weighted adjustment on the input feature map in the channel dimension, and highlight the feature information of important channels; then, based on the feature map enhanced by channel attention, focus on the importance at the spatial position level, perform weighted adjustment on the feature map in the spatial dimension, and further highlight the feature information of important spatial positions. This sequential combination method enables the model to refine and enhance the features from two different dimensions of channels and space, thereby improving the feature expression ability of the model. The data augmentation technology includes geometric transformation (rotation, flipping, scaling), color transformation (brightness adjustment, contrast adjustment, saturation adjustment), and noise addition.

[0025] The channel attention module focuses on the importance of each channel. The channel attention module aims to assign a weight to each channel of the input feature map, thereby emphasizing important channels and suppressing less important channels. The channel attention module mainly includes the following steps: a. Perform global average pooling and global maximum pooling operations on the input feature map. b. Learn the weights for each channel using a fully connected layer (MLP). c. Apply the learned weights to the input feature map to obtain the feature map adjusted by channel attention.

[0026] The spatial attention module focuses on the importance of each spatial position in the feature map. The purpose of the spatial attention module is to assign an attention weight to each position of the input feature map. These weights help the network focus on the regions of interest. The spatial attention module mainly includes the following steps: a. Perform global average pooling and global max pooling operations on the input feature map b. Add the pooled feature maps by channel to obtain two 1D vectors.

[0027] c. Perform a dot product on these two vectors to form an attention weight matrix d. Apply the attention weight matrix to the input feature map to obtain the feature map adjusted by spatial attention.

[0028] Step 3: The Raman spectroscopy detection module automatically moves the placement platform through the Raman interface, performs focusing and dimming operations, and obtains the Raman spectroscopy data of the target cells. The data processing and analysis module processes and analyzes the Raman spectroscopy data, identifies the type and state of the cells, and feeds the results back to the YOLO11 model detection module for verification and correction.

[0029] When the robotic arm module is used for the first time, it is necessary to set the actuator communication port in sequence and perform the conversion between the screen pane coordinate system and the displacement platform plane rectangular coordinate system. To ensure that the microtubule and the probe can accurately follow the mouse for displacement, it is necessary to obtain the mapping relationship between the position of the probe in the microscopic image and the displacement value of the linear actuator. The conversion method is to control the probe to perform a displacement of length s on the x-axis, and record the positions of the linear actuator before and after the displacement as , , and record the positions of the probe tip on the screen pane before and after the displacement as , , then the mapping relationship between the screen pane coordinate system and the displacement platform rectangular coordinate system is:

[0030] Perform the conversion between the screen pane coordinate system and the three-dimensional displacement platform plane rectangular coordinate system:

[0031] In the formula, represents the displacement distance of the linear actuator in the displacement platform plane rectangular coordinate system, represent the coordinates of the probe tip in the screen pane coordinate system;

[0032] The position of the probe tip in the microscopic image can be represented using the screen pane coordinate system; the actual position of the probe is represented using the plane rectangular coordinate system of the displacement platform it is on, and the coordinates of each axis of this coordinate system are the displacement values of the corresponding linear actuator. After determining the mapping relationship between the two coordinate systems, the probe can be precisely controlled using the mouse. It should be noted that this mapping relationship will vary at different magnification factors and needs to be calibrated separately; the instructions input by the host computer to the linear actuator should be absolute positions. If relative displacement instructions are input, it will cause the accumulation of displacement errors and affect the accuracy of control. Since the host computer can control each linear actuator individually, each probe operating arm can move independently or collaboratively, and the system allows any number of probe operating arms to be installed at any position, which can fully meet the operation requirements such as injection, puncture, and transfer of target cells.

[0033] Step Five: According to the final results of Raman spectroscopy detection and YOLO11 model detection, the automated robotic arm operation module will perform operations such as puncturing, transferring, and injecting target cells. Before the operation, the robotic arm will plan the optimal operation path according to the position and state of the target cells. During the operation, the robotic arm will continuously sense its own position and posture and adjust according to the actual situation. At the same time, the robotic arm will precisely control the force and depth of actions such as puncturing, transferring, and injecting to ensure the accuracy and safety of the operation. For example, when performing cell puncture operations, the robotic arm will adjust the force and depth of the puncture needle according to the size and hardness of the cell to avoid excessive damage to the cell.

[0034] In one embodiment, a cell recognition system based on visible light imaging and Raman spectroscopy detection proposed by the present invention includes: a cell imaging and Raman spectroscopy detection module, an automated operation module, a machine learning algorithm module, and a user interface; The cell imaging and Raman spectroscopy detection module includes a confocal Raman microscope and a cell recognition system; The automated operation module includes a host computer and a user operation interaction program, a displacement platform, a linear actuator and its controller, a capillary microtube, a probe, and a fixture; The machine learning algorithm module includes using an improved YOLO11 model to detect target cells; The user interface is used for the user to control the system and display the results.

[0035] a cell imaging and Raman spectroscopy detection module, an automated operation module, a machine learning algorithm module, and a user interface; The cell imaging and Raman spectroscopy detection module includes a confocal Raman microscope and a cell recognition system; The automated operation module includes a host computer and a user operation interaction program, a displacement platform, a linear actuator and its controller, a capillary microtube, a probe, and a fixture; The machine learning algorithm module includes detecting target cells using an improved YOLO11 model; The user interface is used for the user to control the system and display the results.

[0036] Specifically, as Figures 4 to 7 shown, the cell recognition system in the embodiments of the present invention includes a control system, a displacement platform, a confocal Raman microscope, a fixture, a microtube, a probe, a sample fixing table, and a connecting member.

[0037] It should be noted that the microtube or the probe is fixed on the displacement platform through the fixture connecting member. The microtube is made of hollow glass material, one end is connected to a negative pressure device, and the other end is used to aspirate cells. The probe can perform injection or puncture operations on the cells.

[0038] Raman spectroscopy detection can automatically control the movement, focusing, and light adjustment operations of the placement platform, realizing the automation of Raman spectroscopy detection.

[0039] The automated robotic arm operation module in the control system uses high-precision sensors and a control system, can sense its own position and posture in real time, and make precise adjustments according to the position and state of the target cells, precisely controlling the force and depth of the operation actions.

[0040] The data processing and analysis module in the control system preprocesses the Raman spectroscopy data, including operations such as denoising and baseline correction, and analyzes the processed spectroscopy data using machine learning algorithms to identify the type and state of the cells.

[0041] The automated operation module includes a host computer and the user operation interaction program carried thereon, a displacement platform, a linear actuator and its controller, a capillary microtube, a probe fixture. The fixture is connected to the three-dimensional displacement platform through a metal cantilever to form an operating arm. The manual operating rod of the three-dimensional displacement platform is replaced with a linear actuator, and the linear motion output by the actuator is transmitted to the three-dimensional displacement platform to achieve nanoscale operations.

[0042] In the cell recognition system based on visible light imaging and Raman spectroscopy detection of the present invention, a convolutional neural network model is constructed to preprocess the microscopic image, quickly identify the target cells and locate them; a confocal Raman microscope is used to obtain the cell spectrum, and the support vector machine algorithm is combined for metabolic feature verification; the robotic arm in the system can automatically complete cell sorting, culturing, or labeling operations according to the analysis results, and the detection sensitivity is improved through a multimodal data fusion algorithm, and the processing speed is greatly improved compared with traditional methods.

[0043] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0044] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cell identification method based on visible light imaging and Raman spectroscopy detection, characterized in that: include: Obtain a target cell sample, place the target cell sample on a storage table, and install a probe and a capillary microtube of a robotic arm; Connecting the host computer to the microscope camera to obtain the real-time display image of the target cell sample, performing target detection on the real-time display image based on the improved YOLO11 model, and determining the target cell detection result; Connect the host computer to the displacement platform interface, move the stage and perform focusing and dimming operations, obtain Raman spectrum data of the target cell, process and analyze the Raman spectrum data to obtain a cell recognition result, and correct the cell recognition result based on the YOLO11 model to obtain the final cell recognition result; Set the linear actuator communication port in sequence and convert the screen pane coordinate system to the displacement platform plane rectangular coordinate system; According to the Raman spectrum data and the final cell identification result, the operation path planning of the robot arm operation module is carried out to obtain the optimal operation path to achieve the preset operation on the target cell.

2. The cell identification method based on visible light imaging and Raman spectroscopy detection according to claim 1, characterized in that: Connecting the host computer to the microscope camera to obtain a real-time display image of the target cell sample, including: Install Python on the host computer to realize remote control of the storage table and control the storage table to move in the preset direction; After the stage is translated, the microscope camera automatically focuses and collects real-time display images and uploads them to the host computer.

3. The cell identification method based on visible light imaging and Raman spectroscopy detection according to claim 2, characterized in that: Performing target detection on the real-time displayed image based on the improved YOLO11 model to determine the target cell detection result includes: Add CBAM attention mechanism to the YOLO11 model and build an improved YOLO11 model; The CBAM attention mechanism includes a channel attention module and a spatial attention module; The improved YOLO11 model is used to filter the detection boxes below the preset confidence level, mark the remaining detection boxes as target detection boxes, obtain the position information of the target detection boxes, and obtain the target cell detection results.

4. The cell identification method based on visible light imaging and Raman spectroscopy detection according to claim 3, characterized in that: The channel attention module includes the following steps: Perform global average pooling and global maximum pooling operations on the input feature map; Use a fully connected multi-layer perceptron (MLP) to learn the weight of each channel; Apply the learned weights to the input feature map to obtain the channel attention adjusted feature map; Correspondingly, the spatial attention module includes the following steps: Perform global average pooling and global maximum pooling operations on the input feature map; Add the pooled feature maps by channel to get two 1-dimensional vectors; Perform dot product of two 1D vectors to form an attention weight matrix; Apply the attention weight matrix to the input feature map to obtain the spatial attention adjusted feature map.

5. The cell identification method based on visible light imaging and Raman spectroscopy detection according to claim 1, characterized in that: Set the linear actuator communication port in sequence and convert the screen pane coordinate system to the displacement platform plane rectangular coordinate system, including: Control the probe to perform a preset displacement on the x-axis, and obtain the actuator coordinates before and after the displacement of the linear actuator corresponding to the displacement; Obtaining the probe coordinates before and after the displacement of the probe tip on the screen pane; The mapping relationship between the screen pane coordinate system and the displacement platform plane rectangular coordinate system is obtained from the probe coordinates before and after the movement and the preset displacement; determining a rotation angle of coordinates of the probe tip in the screen pane coordinate system based on the probe coordinates before and after the movement; The displacement distance of the linear actuator in the plane rectangular coordinate system of the displacement platform is calculated based on the sine and cosine values ​​of the rotation angle, the probe coordinates before movement, the probe coordinates after movement, the actuator coordinates before movement and a mapping relationship.

6. The cell identification method based on visible light imaging and Raman spectroscopy detection according to claim 1, characterized in that: The preset operations include puncture, transfer and injection.

7. A cell recognition system based on visible light imaging and Raman spectroscopy detection, characterized in that: The cell recognition method based on visible light imaging and Raman spectroscopy detection according to any one of claims 1 to 6 comprises: Cell imaging and Raman spectroscopy detection module, automated operation module, machine learning algorithm module and user interaction interface; The cell imaging and Raman spectroscopy detection module includes a confocal Raman microscope and a cell recognition system; The automated operation module includes a host computer and user operation interactive program, a displacement platform, a linear actuator and its controller, a capillary microtube, a probe and a fixture; The machine learning algorithm module includes the use of an improved YOLO11 model to detect target cells; The user interaction interface is used for users to control the system and display results.

8. The cell recognition system based on visible light imaging and Raman spectroscopy detection according to claim 7, characterized in that: The fixture is connected to the three-dimensional displacement platform through a metal cantilever to form an operating arm. A linear actuator is used to replace the manual operating rod of the three-dimensional displacement platform, and the linear motion output by the actuator is transmitted to the three-dimensional displacement platform to achieve nanometer-level operation.

9. The cell identification system based on visible light imaging and Raman spectroscopy detection according to claim 7, characterized in that: The cell recognition system is used to automatically control the movement, focusing and dimming operations of the displacement platform.

10. The cell identification system based on visible light imaging and Raman spectroscopy detection according to claim 7, characterized in that: The machine learning algorithm module is used to pre-process the Raman spectral data output by the cell recognition system, and use the machine learning algorithm to analyze the processed spectral data to identify the type and state of the cells.

Images