Intelligent laboratory mouse tracheal intubation operation system and method based on mechanical arm linkage

Abstract

The invention discloses an intelligent laboratory mouse tracheal intubation operation system and method based on mechanical arm linkage, and relates to the technical field of technical biomedices.The system comprises a respiratory anaesthesia machine, an anesthesia maintaining box and a mechanical arm, the anesthesia maintaining box is connected with the respiratory anaesthesia machine, and a mechanical arm moving track is arranged on the upper portion in the anesthesia maintaining box; a drawable laboratory mouse device plate for fixing a laboratory mouse is arranged in the anesthesia maintaining box; an operation port of the mechanical arm is provided with a tracheal catheter release device which is used for clamping a tracheal catheter to move to the deep layer in the laryngeal cavity after the oral cavity of the experimental mouse is pried open, and inserting a tracheal cannula into the trachea of the experimental mouse; an image acquisition module for acquiring in-vitro continuous images and in-oral continuous images in real time is arranged beside the tracheal catheter; an image processing module is arranged in the mechanical arm and used for analyzing continuous images in the oral cavity, recognizing and marking the trachea and calculating the time difference of opening and closing of a trachea pipeline window. On the basis of mechanical arm linkage, full-automatic implementation of the tracheal intubation operation in narrow and small lacuna is achieved.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technologies, and particularly to an intelligent tracheal intubation surgical system and method for experimental mice based on robotic arm linkage. Background Art

[0002] Currently, during animal experiments, tracheal intubation of large / small / nude mice is a necessary means. Tracheal intubation is a method of inserting a specially designed thin-diameter flexible tube deep into the larynx of large / small / nude mice and further inserting it into the airway of large / small / nude mice, which can ensure unobstructed respiratory tract, sufficient oxygen supply for large / small / nude mice during the experiment, and prevent various complications, providing optimal conditions for the experiment. At the same time, the rapid development of artificial intelligence (AI) has promoted breakthrough progress in various projects in multiple fields. The deep learning CNN technology is widely used in the field of medical imaging. Currently, some artificial intelligence technologies have been applied to the real medical field, with high diagnostic accuracy. The application of AI has brought more new possibilities for the operation of experimenters and the injury and prognosis of animals.

[0003] Manual tracheal intubation operations in the laboratory have a relatively high risk of failure. Operational errors or failures in tracheal intubation may lead to complications, including mucosal bleeding, laryngeal spasm, bronchospasm, pulmonary aspiration, arrhythmia, laryngeal edema, respiratory tract inflammation, cardiac arrest, and death, causing catastrophic consequences to large / small / nude mice. The success of tracheal intubation is affected by various factors, including the selection of instruments, the experience of the intubator, and the individual conditions of the animals. At present, most tracheal intubations of large / small / nude mice are manually operated by experimenters. Operational errors or immaturity may lead to damage to the tracheal mucosa or the occurrence of some complications, and the operation consumes time and manpower. It is not friendly to inexperienced operators, prone to airway injury or causing local infection and complications after surgery, affecting data stability and unable to conduct continuous research.

[0004] The tracheal intubation technique for the human body has been explored by many researchers to find a device that is simpler and more effective to operate using artificial intelligence (AI) technology. In the literature Patrini I, Ruperti M, Moccia S, Mattos LS, Frontoni E, De Momi E. Transfer learning for informative-frame selection in laryngoscopic videos through learned features. Med Biol Eng Comput. 2020 Jun;58(6):1225-1238. doi: 10.1007 / s11517-020-02127-7. Epub 2020 Mar 24. PMID: 32212052, Patrini et al. developed a method for classifying laryngoscopic video frames based on deep learning. The model can automatically select informative laryngoscopic video frames to help doctors reduce the amount of diagnostic data to be processed. In the literature Wang YY, Hamad AS, Lever TE, Bunyak F. Orthogonal Region Selection Network for Laryngeal Closure Detection in Laryngoscopy Videos. Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:2167-2172. doi: 10.1109 / EMBC44109.2020.9176149. PMID: 33018436, Wang et al. used a convolutional neural network to detect laryngeal adductor reflex events in laryngoscopic videos. In this case, the citation of the AI visual-assisted tracheal intubation technique for large / small / naked mice has become an effective solution. Currently, there is still a lack of a fully automated implementation plan for tracheal intubation surgery based on the recognition of the narrow cavities of experimental mice. Summary of the Invention

[0005] The object of the present invention is to provide an intelligent tracheal intubation surgery system and method for experimental mice based on robotic arm linkage. Through the combination of robotic arm linkage and artificial intelligence technology, the recognition of the narrow cavities of experimental mice realizes the automated progress of tracheal intubation surgery.

[0006] To achieve the above object, the present invention provides the following solutions: An intelligent tracheal intubation surgical system for experimental mice based on robotic arm linkage, comprising: a respiratory anesthetic machine, an anesthesia maintenance box, and a robotic arm. The anesthesia maintenance box is connected to the respiratory anesthetic machine through a rigid anesthesia gas connection pipe. A robotic arm moving track is arranged above the anesthesia maintenance box. An extractable experimental mouse carrier plate for placing the experimental mouse and an experimental mouse fixing device for fixing the four limbs of the experimental mouse are arranged inside the anesthesia maintenance box; An endotracheal tube releasing device is arranged at the operation port of the robotic arm. The endotracheal tube releasing device is used for prying open the mouth of the experimental mouse and then clamping the endotracheal tube to move deeper into the laryngeal cavity of the experimental mouse; An image acquisition module is arranged beside the endotracheal tube, and an image processing module is arranged inside the robotic arm. The image acquisition module is used for continuously acquiring real-time external images and intraoral images of the experimental mouse and transmitting them to the image processing module; The image processing module is used for automatically analyzing the intraoral continuous images according to the image data, identifying and marking the trachea of the experimental mouse, and then automatically calculating the time difference of the opening and closing of the tracheal tube window of the experimental mouse. The robotic arm adjusts the position of the endotracheal tube according to the position of the trachea of the experimental mouse, automatically pauses moving after determining the optimal tracheal intubation position, and inserts the tracheal intubation into the trachea of the experimental mouse by using the endotracheal tube releasing device.

[0007] Further, the image processing module includes a preprocessing unit, a convolutional neural network unit, and a Python algorithm unit. The Python algorithm unit is used for calculating the time difference of the opening and closing of the tracheal tube window of the experimental mouse. The convolutional neural network unit includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer; The preprocessing unit preprocesses the intraoral continuous images to obtain a picture pixel matrix. The input layer is used for receiving three consecutive frames of continuous images, extracting five-channel information of the gray scale, horizontal coordinate gradient, vertical coordinate gradient, x optical flow, and y optical flow of each frame. The convolutional layer is used for identifying the spatial pattern in the picture pixel matrix, performing a convolutional operation on the 3D convolutional kernel of the five channels to obtain features. The pooling layer is used for reducing the size of the matrix to reduce the parameters in the entire neural network. The fully connected layer is used for classifying the features to obtain a classification result and transmitting it to the output layer. The output layer is used for outputting the classification result in S4.4 to realize the identification and marking of the trachea.

[0008] Further, the robotic arm is a multi-axis robotic arm. The robotic arm moving track includes a horizontal track and a vertical track. The image acquisition module is a lens, and the lens is equipped with an active lighting device. The tracheal intubation is a balloon tracheal intubation.

[0009] Further, the tracheal catheter release device is a scissor-type tracheal catheter release device. A tiny spring is placed at the center of the scissor-type tracheal catheter release device. The scissor-type tracheal catheter release device holds the tracheal catheter, making it located at the center of the tracheal catheter release device and connected to the spring. The tracheal catheter connected to the scissor-type tracheal catheter release device tightly wraps the balloon trachea; the tracheal catheter is a symmetric rigid semi-tube.

[0010] The present invention also provides a method for intelligent tracheal intubation surgery of experimental mice based on robotic arm linkage, which is applied to the intelligent tracheal intubation surgery system of experimental mice based on robotic arm linkage, and includes the following steps: S1. Turn on the breathing anesthesia machine, and the breathing anesthesia machine continuously injects isoflurane gas into the anesthesia maintenance box to ensure the anesthesia state of the experimental mouse until the muscle reflex of the experimental mouse disappears; S2. Fix the four limbs of the anesthetized experimental mouse to the retractable experimental mouse carrier board through the experimental mouse fixing device; S3. Start the intubation operation: Collect continuous external images of the experimental mouse through the image acquisition module, judge whether the tracheal catheter release device reaches the oral position of the experimental mouse according to the continuous external images, pry open the mouth of the experimental mouse by the tracheal catheter release device to completely expose the larynx, and then clamp the tracheal intubation and move it into the laryngeal cavity of the experimental mouse; S4. After inserting into the oral cavity, the image acquisition module continues to collect continuous images in the oral cavity of the experimental mouse, and then uses the image processing module to preprocess the continuous images in the oral cavity, and based on the convolutional neural network, identify and label the trachea of the experimental mouse, and calculate the time difference of the opening and closing of the tracheal duct window of the experimental mouse; S5. The robotic arm adjusts the position of the tracheal catheter according to the position of the trachea of the experimental mouse. After determining the optimal tracheal intubation position, it automatically pauses the movement, and uses the tracheal catheter release device to insert the tracheal intubation into the trachea of the mouse; S6. Use the robotic arm to take out the instruments in the oral cavity of the experimental mouse except the tracheal intubation, manually take out the experimental mouse from the anesthesia maintenance box, and then the breathing anesthesia machine automatically stops releasing isoflurane gas to complete the whole process of tracheal intubation of the experimental mouse.

[0011] Further, in the step S4, the identification and labeling of the trachea of the experimental mouse are realized based on the convolutional neural network, and specifically include the following steps: S4.1. Preprocess the continuous images in the oral cavity through the preprocessing unit in the image processing module to obtain a picture pixel matrix; S4.2. The input layer receives continuous images of 3 consecutive frames, and extracts five-channel information of the gray scale, horizontal coordinate gradient, vertical coordinate gradient, x optical flow, and y optical flow of each frame; S4.3. Identify the spatial patterns in the picture pixel matrix by the convolutional layer, perform convolution operations on the 3D convolutional kernels of the five channels to obtain features; S4.4. Reduce the size of the picture pixel matrix by the pooling layer to reduce the parameters in the entire neural network; repeat steps S4.3 and S4.4; S4.5. Classify the features by the fully connected layer to obtain a classification result and transmit it to the output layer; S4.6. Finally, output the classification result in S4.4 by the output layer to achieve the identification and annotation of the trachea.

[0012] Further, in S4, calculating the time difference of the opening and closing of the tracheal duct window of the experimental mouse specifically includes: After the preprocessing unit performs image recognition, it extracts the images of the airway window opening and closing and transmits them to the Python algorithm unit in the image processing module. The Python algorithm unit automatically calculates the time difference of the tracheal opening and closing of the experimental mouse according to the time difference between the transmitted images.

[0013] Further, the movement principle of the robotic arm is as follows: Grayscale the recorded images, calculate the central position of the robotic arm movement, establish a three-dimensional coordinate axis based on the central position. This three-dimensional coordinate axis is used for trajectory planning and motion planning. Use the coordinates [x, y, z] to represent the spatial position points within the movable range of the robotic arm, ensure that the operating port of the robotic arm can reach each spatial position point in the anesthesia maintenance box, and use a visual vision device to ensure the accurate positioning and visual tracking of the robotic arm for the experimental mouse.

[0014] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects: The intelligent tracheal intubation surgery system and method for experimental mice based on the linkage of the robotic arm provided by the present invention dynamically identify and label the position of the tracheal duct in the laryngeal cavity of the experimental mouse and the opening time of the tracheal duct through deep learning CNN (CNN: Convolutional Neural Network), and at the same time link the robotic arm to realize the dynamic insertion of tracheal intubation, achieving fully automated tracheal intubation. The present invention has the following advantages: 1. High accuracy: The artificial intelligence (AI)-assisted tracheal intubation technology uses deep learning algorithms to visually analyze the position, composition, width, and the opening time of the tracheal duct opening window of the digestive tract, which is beneficial to the correct insertion of tracheal intubation.

[0015] 2. Time and labor cost savings: Using artificial intelligence (AI) technology to achieve full automation of tracheal intubation only takes 5s - 7s to complete the experiment. It can reach the experimental results of correct tracheal intubation for a large number of large / small / nude mice with single-person control of the equipment, greatly reducing the complexity of tracheal intubation technology and saving a large amount of time and labor costs.

[0016] 3. Risk rate reduction: The experimental results show that compared with the artificial tracheal intubation experiment, due to different artificial methods, the risk rate of the tracheal intubation experiment operated by this method is only 0.5% - 1%. The airways of large / small / nude mice are very narrow. Using deep learning algorithms and combined with the application in the field of vision is conducive to finding the intubation tube more accurately, with less damage and avoiding the occurrence of complications caused by various mistakes.

[0017] 4. High stability: This method establishes a full automation model. The artificial tracheal intubation method has obvious structural errors due to different personnel levels, while this method using artificial intelligence technology has high full automation stability.

[0018] 5. Intelligent recognition of narrow cavities: The image recognition of the present invention is different from the visual field recognition of conventional methods. The tracheas of large / small / nude mice are extremely thin and small, and the visible range is narrow. It is not the observation of a large visual field where the observed object can be directly observed from a distance and the object can be gradually magnified visually when approaching the observed object. The visual observation of a narrow tube will quickly present the measured target without a visual buffering process. This technology realizes the intelligent recognition of narrow cavities through deep learning algorithms, and realizes the process of quickly recognizing and classifying the observed objects that appear quickly.

[0019] 6. Automatic technology: The present invention uses a chip to automatically control the movement of the robotic arm to perform automatic surgery on narrow cavities, without the need for full-process manual operation, saving labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0021] Figure 1 It is a structural diagram of the intelligent tracheal intubation surgery system for experimental mice based on robotic arm linkage of the present invention; Figure 2 It is a flowchart of the intelligent tracheal intubation surgery method for experimental mice based on robotic arm linkage in the embodiment of the present invention; Figure 3 It is a schematic diagram of the intelligent tracheal intubation surgery method for experimental mice based on robotic arm linkage in the embodiment of the present invention; Figure 4 It is a schematic diagram of the laryngeal cavity structure of a mouse. Among them, A is the oral cavity wall, B is the trachea, C is the epiglottis, D is the esophagus, and E is the tracheal intubation; Explanation of the reference numerals in the drawings: 1. Anesthesia ventilator; 2. Hard anesthesia gas connection pipeline; 3. Hard oxygen connection pipeline; 4. Pull-out experimental mouse carrier plate; 5. Experimental mouse fixing device; 6. Soft oxygen connection pipeline; 7. Horizontal track; 8. Vertical track; 9. Robotic arm; 10. Robotic arm operation port; 11. Scissor-type tracheal catheter release device; 12. Tracheal catheter; 13. Image acquisition module (lens); 14. Guide wire; 15. Balloon tracheal intubation; 16. Balloon inflation connection tube; 17. Interface between the hose and each part of the tracheal intubation pipeline; 18. Anesthesia maintenance box. Specific implementation manners

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] The purpose of the present invention is to provide an intelligent tracheal intubation surgery system and method for experimental mice based on robotic arm linkage. Through the combination of robotic arm linkage and artificial intelligence technology, the recognition of the narrow cavity of the experimental mouse is realized, and the automation of the tracheal intubation surgery is achieved.

[0024] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the drawings and specific implementation manners.

[0025] Embodiment 1 As Figure 1 shown, the embodiment of the present invention provides an intelligent tracheal intubation surgery system for experimental mice based on robotic arm linkage, including: an anesthesia ventilator 1, an anesthesia maintenance box 18, and a robotic arm 9; The anesthesia maintenance box 18 is used to maintain a certain concentration of anesthesia gas to keep small animals in an anesthetized state; The anesthesia maintenance box 18 is connected to the respiratory anesthesia machine 1 through a hard anesthesia gas connection pipeline 2. A robotic arm moving track is arranged above the anesthesia maintenance box 18. The robotic arm moving track includes a horizontal track 7 and a vertical track 8; a pull-out experimental mouse carrier plate 4 for placing experimental mice and an experimental mouse fixing device 5 for fixing the limbs of the experimental mice are arranged inside the anesthesia maintenance box 18; The operation port 10 of the robotic arm 9 is provided with a tracheal catheter release device 11, and the tracheal catheter release device 11 is used to pry open the mouth of the experimental mouse and then clamp the tracheal catheter 12 to move deeper into the laryngeal cavity of the experimental mouse; An image acquisition module 13 is arranged beside the tracheal catheter 12, and an image processing module is arranged inside the robotic arm 9; the image acquisition module is used to collect the continuous external images and the continuous oral images of the experimental mouse in real time and transmit them to the image processing module; the image processing module is used to automatically analyze the continuous oral images according to the image data, identify and label the trachea of the experimental mouse, and then automatically calculate the time difference between the opening and closing of the tracheal duct window of the experimental mouse; the robotic arm adjusts the position of the tracheal catheter 12 according to the position of the trachea of the experimental mouse, automatically pauses at an appropriate angle, and uses the tracheal catheter release device to insert the tracheal intubation 15 into the trachea of the mouse.

[0026] In this embodiment, the robotic arm 9 is connected to the robotic arm moving track through gears, and the controlled port of the robotic arm 9 is connected to the balloon tracheal intubation sleeve, the endoscope and the tracheal intubation release device.

[0027] In this embodiment, the image processing module is a chip, and the chip is built into the robotic arm 9.

[0028] In this embodiment, the image processing module includes a preprocessing unit, a convolutional neural network unit, and a Python algorithm unit; the Python algorithm unit is used to calculate the time difference between the opening and closing of the tracheal duct window of the experimental mouse; the convolutional neural network unit includes an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer; The preprocessing unit preprocesses the continuous oral images to obtain a picture pixel matrix; the input layer is used to receive three consecutive frames of continuous images and extract five-channel information of the gray scale, the horizontal coordinate gradient, the vertical coordinate gradient, the x optical flow, and the y optical flow of each frame; the convolutional layer is used to identify the spatial patterns in the picture pixel matrix to obtain features; the pooling layer is used to reduce the size of the matrix to reduce the parameters in the entire neural network; repeat steps S4.3 and S4.4; the fully connected layer is used to classify the features to obtain a classification result and transmit it to the output layer; the output layer is used to output the classification result in S4.4 to realize the identification and labeling of the trachea.

[0029] In this embodiment, the robotic arm is a multi-axis robotic arm; the image acquisition module is an endoscope, and the endoscope is equipped with an active lighting device; the diameter of the endoscope lens is small, the images taken are clear, and the lens does not heat up, and the images can be output to the robotic arm artificial intelligence (AI) chipset.

[0030] In this embodiment, the tracheal catheter releasing device 11 is a scissor-type tracheal catheter releasing device. A tiny spring is placed at the center of the scissor-type tracheal catheter releasing device. The scissor-type tracheal catheter releasing device holds the tracheal catheter 12, making it located at the center of the tracheal catheter releasing device and connected to the spring. The tracheal catheter 12 connected to the scissor-type tracheal catheter releasing device tightly wraps the balloon trachea. The tracheal catheter 12 is a symmetric rigid semi-tube. In a further embodiment, the scissor-type tracheal catheter releasing device can be a scissor-type iron clip.

[0031] In this embodiment, the tracheal intubation 15 is a balloon tracheal intubation cannula, that is, there is a balloon at the front end of the tracheal intubation, and this balloon is connected to a pressure device, enabling the inflation and deflation of the balloon. Specifically: The balloon is connected to the oxygen rigid communication pipeline through a balloon inflation connecting tube, and the oxygen rigid communication pipeline is connected to the anesthesia breathing integrated machine, thereby realizing the inflation and deflation of the balloon.

[0032] In this embodiment, the system further includes an oxygen flexible communication pipeline 6, which can be connected to the anesthesia ventilator 1 and the tracheal intubation 15 through a hose at the interface 17 between each part of the tracheal intubation, filling the tracheal intubation with anesthetic gas. After the tracheal intubation 15 is inserted into the mouth of the experimental mouse, it leaves the anesthesia maintenance box 18 for subsequent surgical operations.

[0033] In a further embodiment, a guide wire 14 is included in the balloon tracheal intubation, facilitating the insertion of the trachea into the tracheal window of the mouse.

[0034] The physical properties of the system are as follows: Applicable animal range: including all large / small / nude mice; Applicable narrow cavity range: the respiratory tracts of all large / small / nude mice; Volume of the anesthesia maintenance box: 40*40*80 cm Endoscope resolution: 720P, 24FPS, lens diameter is 2.8 mm; Robotic arm: multi-axis linkage, arm span 350 mm, effective grasping range: radius ≤ 30 cm, the area of a semi-circle centered on the central axis, repeat positioning accuracy ±0.5 mm; Artificial intelligence (AI): The microprocessor uses Broadcom BCM2711 64-bit 1.5 GHz quad-core + VideoCore VI @ 500 MHz, artificial intelligence (AI) computing power: 200 GFLOPS, operating system: Ubuntu mate 20.04 LTS + ROS Noetic, programming language: Python.

[0035] Embodiment 2 As Figures 2-3As shown in the figure, an intelligent tracheal intubation surgery method for experimental mice based on robotic arm linkage provided by an embodiment of the present invention is applied to the intelligent tracheal intubation surgery system for experimental mice based on robotic arm linkage described in Embodiment 1, and includes the following steps: S1. Turn on the breathing anesthesia machine, and the breathing anesthesia machine continuously injects isoflurane gas into the anesthesia maintenance box to ensure the anesthesia state of the experimental mouse until the muscle reflex of the experimental mouse disappears; S2. Manually draw out the experimental mouse fixing plate from the anesthesia maintenance box, fix the limbs of the anesthetized experimental mouse to the pull-out experimental mouse carrier plate through the experimental mouse fixing device, and return the pull-out experimental mouse carrier plate to its original position; S3. Start the intubation operation: Collect continuous external images of the experimental mouse through the image acquisition module, and judge whether the tracheal catheter release device reaches the oral position of the experimental mouse according to the continuous external images. When it reaches the oral position, the tracheal catheter release device pries open the mouth of the experimental mouse to completely expose the larynx, and then clamps the tracheal intubation and moves it into the laryngeal cavity of the large / small / naked mouse. This process is to control the insertion of the intubation trachea into the mouth of the experimental mouse by automatically adjusting the angles and lengths between the axes of the robotic arm; S4. After inserting into the mouth, the image acquisition module continues to collect continuous images inside the mouth of the experimental mouse, and then uses the image processing module to preprocess the continuous images inside the mouth, and based on the convolutional neural network, identify and label the position of the trachea of the experimental mouse, and calculate the time difference between the opening and closing of the tracheal duct window of the experimental mouse; S5. The robotic arm adjusts the position of the tracheal catheter according to the position of the trachea of the experimental mouse. After determining the optimal tracheal intubation position, it automatically pauses the movement, and uses the tracheal catheter release device to insert the tracheal intubation into the trachea of the mouse; S6. Use the robotic arm to take out the instruments in the mouth of the experimental mouse except the tracheal intubation, manually take out the experimental mouse from the anesthesia maintenance box, and then the breathing anesthesia machine automatically stops releasing isoflurane gas to complete the whole process of tracheal intubation of the experimental mouse.

[0036] In this embodiment, in S4, the identification and labeling of the trachea of the experimental mouse are realized based on the convolutional neural network, which specifically includes the following steps: S4.1. Preprocess the continuous images inside the mouth through the preprocessing unit in the image processing module to obtain a picture pixel matrix; S4.2. The input layer receives continuous images of 3 consecutive frames, and extracts five-channel information of the gray scale, horizontal coordinate gradient, vertical coordinate gradient, x optical flow, and y optical flow of each frame; S4.3. The convolutional layer identifies the spatial pattern in the picture pixel matrix, performs a convolution operation on a 3D convolution kernel of 3*3*3 for the five channels, and obtains features by extracting the information in the image pixel matrix; the information in the image pixel matrix includes local features of the image, such as edges, textures, corner points, etc.

[0037] S4.4. Reduce the size of the image pixel matrix by the pooling layer to reduce the parameters in the entire neural network; repeat steps S4.3 and S4.4 to increase the number of images; S4.5. Classify the features by the fully connected layer, obtain the classification result and transmit it to the output layer; S4.6. Finally, the output layer outputs the classification result in S4.4 to realize the recognition and annotation of the trachea; In the above process, the 3D-CNN algorithm is used to capture the three-dimensional features of the dynamic image. Since the 3D-CNN can make full use of the convolutional kernel during feature extraction, the computational amount and storage amount of model construction are greatly reduced, and the configuration requirements of the construction equipment are reduced. In this embodiment, the ResNet-50 residual structure can also be used to solve the problems of the decline in classification performance, the slowdown of network convergence speed and the decline in accuracy rate caused by the increase in the depth of the conventional CNN model; the softmax function is used for construction to realize the task of correct classification and recognition of continuous images.

[0038] In this embodiment, in S4, calculating the time difference of the opening and closing of the tracheal duct window of the experimental mouse specifically includes: After the preprocessing unit performs image recognition, the images of the airway window opening and closing are extracted and transmitted to the Python algorithm unit in the image processing module. The Python algorithm unit automatically calculates the time difference of the tracheal opening and closing of the experimental mouse according to the time difference between the transmitted images, and repeats the experiment multiple times.

[0039] In this embodiment, the movement principle of the robotic arm is as follows: The recorded images are grayscale processed, the central position of the robotic arm movement is calculated, and a three-dimensional coordinate system is established according to the central position. This three-dimensional coordinate system is used for trajectory planning and motion planning. The spatial position points within the movable range of the robotic arm are represented by the coordinates [x, y, z], ensuring that the operating ports of the robotic arm can reach each spatial position point in the anesthesia maintenance box. A visual vision device is used to ensure the accurate positioning and visual tracking of the robotic arm for the experimental mouse. After having a preliminary understanding of the model, the model is used to directly and automatically insert a tracheal intubation into the experimental mouse body.

[0040] This embodiment is screened according to the following criteria to verify the feasibility of the present invention: 180 rats, 180 mice, and 180 nude mice are respectively selected (the animals selected for the experiment comply with ethics, number: SQ2024143). It is ensured that the weight, age, and physical signs of each group of large / small / nude mice are different. Each group of mice is divided into three groups of high, medium, and low weights according to body weight, with 60 mice in each group (each mouse is trained 10 times). Training group: 45 mice, verification group: 15 mice.

[0041] Artificially insert tracheal intubation into some of the large / small / nude mice in the training group, record the dynamic images in the visual camera, perform data processing, and use the conventional Otsu threshold to automatically segment the tissues in the real-time dynamic images [1]. The tissue area of each frame of image is cut into non-overlapping 250-micron square small pieces, magnified 25 times, and the size is adjusted to 224×224 pixels. This size is determined based on experience. Experienced clinicians manually excluded the non-airway images, and manually marked the obtained images, including the secretions and anatomical structures of the large and small mice, etc., so that the error rate of the model automatically identifying the trachea is almost 0. The labeled images are saved in different data sets in PNG format. The data is divided into a training set, a validation set, and a test set.

[0042] Manually analyze the mistakes or errors of the model during tracheal intubation and modify them until all the large / small / nude mice in the training group are trained. Then, randomly select 20 large / small / nude mice from the training group of large / small / nude mice for a second experiment of tracheal intubation to analyze the accuracy of this method. Finally, use another 15 large / small / nude mice for external verification to prove the precision and timeliness of this method.

[0043] Example 3 The embodiment of the present invention is a specific case of the intelligent tracheal intubation surgery method for experimental mice based on the linkage of robotic arms described in Example 2.

[0044] In this embodiment, mice are randomly selected. The mice (the animals selected for the experiment comply with ethics, North China University of Science and Technology Ethics Number: SQ2024143), at a room temperature of 24°C, a relative humidity of 55%, with air circulation, and the experimental research is carried out under natural light.

[0045] The operator places the anesthetized mouse on a self-made pullable device board for large and small mice and fixes the four limbs of the mouse.

[0046] Turn on the RWD (Rewod) customized respiratory anesthesia integrated machine. The respiratory anesthesia machine continuously injects isoflurane gas into the RWD (Rewod) customized anesthesia maintenance box to ensure the anesthetized state of the mouse. After the muscle reflex of the mouse disappears, the experiment begins.

[0047] Using the method of the present invention, when starting up, the device performs automated operations to smoothly insert the balloon tracheal intubation into the laryngeal cavity of a mouse. (First, the robotic arm collects image data through a 2.8 mm customized lens. The built-in STM32 chip uses artificial intelligence (AI) algorithms to identify and calculate the position of the mouse's mouth. The customized DOFBOT robotic arm (main board: Raspberry Pi 4B, computing power: 0.2 TFLOPS, CPU: Quad-Core Arm Cortex-A72 1.5 GHz, GPU: Broadcom VideoCore IV, OLED device) clamps and moves the customized tracheal catheter release device to pry open the mouse's mouth to make it open, completely exposing the larynx. The robotic arm automatically adjusts the angles and lengths between each axis to adapt to the correct insertion of the rat tracheal intubation. After inserting into the mouse's mouth, the lens collects image data inside the mouse's mouth and uses artificial intelligence (AI) to automatically analyze and preprocess the continuous dynamic images. Using CNN (CNN: Convolutional Neural Network), it collects image data inside the mouse's mouth and collects features for deep learning. Through the neural network input of the pixel matrix of the collected picture in the input layer, identifying the spatial patterns in the image in the convolutional layer to obtain features, reducing the size of the matrix in the pooling layer to reduce the parameters in the entire neural network, the classification task of features in the fully connected layer, and the output of the classification results in the output layer, the tracheal recognition and labeling task is achieved. At the same time, according to the collected image data, the time difference of the opening and closing of the mouse tracheal duct window can be automatically calculated. The robotic arm adjusts the position of the tracheal catheter according to the tracheal position, automatically pauses at the appropriate angle, and uses the microscopic tracheal release device to smoothly insert the tracheal intubation into the mouse trachea, making the abdominal undulation frequency of the mouse the same as the gas inflow frequency of the ventilator, and correctly completing the tracheal intubation operation.) Move the robotic arm to remove the instruments in the animal's mouth except the tracheal intubation, and manually take out the mouse from the empty box. The respiratory anesthesia machine automatically stops releasing isoflurane gas, completing the entire process of tracheal intubation.

[0048] The experimental results are as follows: The tracheal intubation experiment was completed in only 5 s. After the experiment, the mouse showed no abnormalities, and its chest undulated regularly with the inflow of the ventilator gas. The mouse had no loose or fallen teeth, no gastric distension, no mucosal bleeding, vocal cord paralysis, laryngeal spasm, bronchospasm, pulmonary aspiration, laryngeal edema, respiratory tract inflammation, arrhythmia, cardiac arrest, or death. Identified by an experimentalist with 5 years of experience in animal tracheal intubation, the tracheal intubation of the mouse was successful without problems or complications.

[0049] Example 4 This embodiment of the present invention is another case of the intelligent tracheal intubation surgery method for experimental mice based on the linkage of the robotic arm described in Example 2.

[0050] Rats were randomly selected (the animals selected for the experiment were ethically compliant, number: SQ2024143) and tested at 24°C room temperature, 55% relative humidity, air circulation, and natural light. The operator placed the anesthetized rat on a self-made retractable rat and mouse device board and fixed the rat's limbs.

[0051] Turn on the RWD (RWD) respiratory anesthesia integrated machine, which continuously injects isoflurane gas into the RWD (RWD) customized anesthesia maintenance box to ensure the anesthesia state of the rat. After the rat's muscle reflex disappears, the experiment begins.

[0052] Using this method, the device is turned on and automatically operated to smoothly insert the airbag endotracheal tube into the rat's laryngeal cavity. (The robotic arm first collects image data through a 2.8mm custom lens, and the built-in STM32 chip uses artificial intelligence (AI) algorithms to identify and calculate the position of the rat's mouth. The robotic arm clamps the mobile clamp to pry open the rat's mouth to expose the throat completely. The DOFBOT robotic arm (mainboard: Raspberry Pi 4B, computing power: 0.2TFLOPS, CPU: Quad-Core Arm Cortex-A72 1.5GHz, GPU: Broadcom VideoCore IV, OLED device) automatically adjusts the angles and lengths between the axes to adapt to the correct insertion of the rat's endotracheal tube. After insertion into the rat's mouth, the lens collects the image data in the rat's mouth and uses artificial intelligence (AI) to automatically analyze and pre-process the continuous dynamic images. Using CNN (CNN: Convolutional Neural Network, Convolutional NeuralNetwork), collects intraoral imaging data of rats and collects features for deep learning. Through the neural network input of the image pixel matrix collected by the input layer, the spatial pattern in the image is recognized by the convolution layer to obtain features, the size of the matrix is ​​reduced by the pooling layer to reduce the parameters in the entire neural network, the classification task of the features of the fully connected layer and the output of the classification results of the output layer, the trachea identification and labeling task is achieved. At the same time, the time difference of the opening and closing of the rat tracheal tube window can be automatically calculated based on the collected image data. The robotic arm adjusts the position of the tracheal tube according to the position of the trachea, automatically pauses at the appropriate angle, and uses the microscopic tracheal release device to smoothly insert the tracheal tube into the rat's trachea, so that the frequency of the rat's abdominal cavity fluctuations and the frequency of the ventilator's gas inflow are the same, and the tracheal intubation operation is correctly completed. ) The mechanical arm was moved to remove all instruments except the tracheal tube from the animal's mouth, and the rat was manually taken out of the empty box. The respiratory anesthesia machine automatically stopped releasing isoflurane gas, completing the entire tracheal intubation process.

[0053] The experimental results are as follows: The experiment was completed in just 4.5 seconds. After the experiment, the rats showed no abnormalities and their chests rose and fell regularly with the inflow of ventilator gas. There were no loose or falling teeth, no stomach swelling, no mucosal bleeding, vocal cord paralysis, laryngeal spasm, bronchospasm, pulmonary aspiration, laryngeal edema, respiratory inflammation, arrhythmia, cardiac arrest, and death. The rats were successfully intubated without problems or complications, according to an experimenter with 5 years of experience in animal intubation.

[0054] Example 5 The embodiment of the present invention is another example of the intelligent tracheal intubation surgical method for experimental mice based on robotic arm linkage described in Example 2.

[0055] Nude mice were randomly selected (animals selected for the experiment were ethically compliant, number: SQ2024143), room temperature of 28°C, maximum daily temperature difference of no more than 4°C, relative humidity of 50%, air circulation, sterile environment, and natural light were used for the experiment. 1. The operator places the anesthetized nude mouse on a self-made retractable rat and mouse device board and fixes the limbs of the nude mouse.

[0056] 2. Turn on the RWD respiratory anesthesia integrated machine, which continuously injects isoflurane gas into the RWD customized anesthesia maintenance box to ensure the anesthesia state of the nude mice. After the muscle reflex of the nude mice disappears, the experiment begins.

[0057] 3. Use this method, turn on the device, and perform automated operation to smoothly insert the airbag endotracheal tube into the laryngeal cavity of the nude mouse. (The robotic arm first collects image data through a 2.8mm customized lens, and the built-in STM32 chip uses artificial intelligence (AI) algorithm to identify and calculate the position of the nude mouse's mouth. The robotic arm clamps the mobile clamp to pry open the nude mouse's mouth to expose the throat completely. The DOFBOT robotic arm (mainboard: Raspberry Pi 4B, computing power: 0.2TFLOPS, CPU: Quad-Core Arm Cortex-A72 1.5GHz, GPU: Broadcom VideoCore IV, OLED device) automatically adjusts the angle and length between each axis to adapt to the correct insertion of the nude mouse endotracheal tube. After insertion into the rat's mouth, the lens collects the image data in the nude mouse's mouth and uses artificial intelligence (AI) to automatically analyze and pre-process the continuous dynamic images. Using CNN (CNN: Convolutional Neural Network, Convolutional NeuralNetwork), collects the image data in the mouth of nude mice and collects features for deep learning. The neural network input of the image pixel matrix is ​​collected by the input layer, the spatial pattern in the image is recognized by the convolution layer to obtain features, the size of the matrix is ​​reduced by the pooling layer to reduce the parameters in the entire neural network, the classification task of the features of the fully connected layer and the output of the classification results of the output layer are completed to achieve the trachea identification and labeling task. At the same time, the time difference of the opening and closing of the rat tracheal tube window can be automatically calculated based on the collected image data. The robotic arm adjusts the position of the tracheal tube according to the position of the trachea, automatically pauses at the appropriate angle, and uses the microscopic tracheal release device to smoothly insert the tracheal tube into the trachea of ​​the nude mouse, so that the frequency of the nude mouse abdominal cavity fluctuations and the frequency of the ventilator gas inflow are the same, and the tracheal intubation operation is correctly completed. ) 4. Move the mechanical arm to remove all instruments except the tracheal tube from the animal's mouth, and manually take the nude mouse out of the empty box. The respiratory anesthesia machine automatically stops releasing isoflurane gas, completing the entire tracheal intubation process.

[0058] The experimental results are as follows: The experiment was completed in just 4.5 seconds. After the experiment, the nude mice showed no abnormalities and the chest cavity rose and fell regularly with the inflow of ventilator gas. There were no loose or falling teeth, no stomach swelling, no mucosal bleeding, vocal cord paralysis, laryngeal spasm, bronchospasm, pulmonary aspiration, laryngeal edema, respiratory inflammation, arrhythmia, cardiac arrest and death. According to the identification of an experimenter with 5 years of experience in animal tracheal intubation, the nude mouse tracheal intubation was successful without any problems or complications.

[0059] Comparative Example Three persons with the same technical level performed manual tracheal intubation on rats. The manual operation and the experiment using this method were carried out simultaneously. Each operator performed tracheal intubation on 50 large / small / nude mice and calculated the time taken and recorded the prognosis.

[0060] Manual operation fully meets the requirements of endotracheal intubation process.

[0061] One patient used blind intubation, one used transcervical intubation, and one used tracheotomy intubation.

[0062] Anesthetized experimental mice were weighed and anesthetized with isoflurane according to their body weight.

[0063] Comparative Example 1 - Blind intubation: The operator fixes the anesthetized experimental mouse in a supine position on a rat fixing board, pulls the upper incisor of the experimental mouse backward to tilt the head back and fix it, holds the tongue of the experimental mouse with toothless forceps in one hand and pulls it upward, stretches the mouse's laryngeal cavity as much as possible to expose the airway opening, and performs tracheal intubation of the experimental mouse. The front end of the tube is against the front diaphragm and inserted at a slightly inclined angle. When there is no obvious resistance and abnormal sensation of tracheal intubation, remove the guide wire 14 in the catheter. The entire operation process needs to be performed carefully and slowly to ensure the accuracy and safety of the intubation operation.

[0064] Comparative Example 2 - Transcervical intubation method: The operator fixes the anesthetized experimental mouse in a supine position on a rat and mouse fixing board, places a high-intensity cold light source 5 cm from the neck of the experimental mouse to illuminate the rat's throat through the neck, uses tweezers to gently pull the experimental mouse's tongue upward to the side of the lower teeth, and presses the tongue downward with a tongue depressor to open the experimental mouse's mouth as wide as possible to expose the airway window. The light source illuminates the rat's throat through the experimental mouse's mouth, and the opening and closing of the experimental mouse's airway opening can be clearly seen. The operator inserts a tracheal tube into the experimental mouse's laryngeal cavity at the appropriate time.

[0065] Comparative Example 3 - This comparative example adopts the tracheotomy and intubation method, and the specific operation is as follows: the operator shaves and disinfects the anesthetized experimental mouse, then fixes the experimental mouse and cuts the skin muscle layer and fascia, carefully frees the thyroid gland and surrounding tissues, and fixes the trachea after touching the trachea with circular cartilage, gently picks open the trachea with a sharp knife until a 1mm incision is formed, and gently inserts the catheter downward along the trachea.

[0066] The method of the present invention: The automated robotic method is similar to the manual method of anesthesia, but uses technology to automate the process.

[0067] The experimental time, experimental damage and subsequent prognosis were recorded.

[0068] 1. Experimental time The average time of the whole process of tracheal intubation for four people was calculated as follows: Tracheotomy intubation method: 20.15±1.2min, blind intubation method: 6.15±1.8min, transcervical intubation method: 4.27±1.5min, and the entire process of AI model tracheal intubation is 0.75±0.2min.

[0069] 2. Experimental injury conditions With the transillumination intubation method through the neck, the tracheal window is not clearly observable, and the intubation success rate is low; The success rate of the blind intubation method is not high, and the tracheal window is very unclear. Only the appearance of the airway window in a single direction can be observed, and the complete original appearance of the airway window cannot be seen. Whether to insert by the blind intubation method needs to be judged by the operator's experience. Judging whether to insert based solely on feeling is not friendly to operators with insufficient experience, and this method is extremely likely to cause airway mucosal injury, the occurrence of airway inflammation, and even suffocation and death of experimental mice; With the tracheotomy intubation method, the incised wound stimulates the experimental mice themselves, resulting in changes in some data of the experimental mice and affecting the experimental results. Other methods except the automatic robot method have varying degrees of injury.

[0070] 3. Subsequent prognosis conditions The proportion of poor prognosis with the blind intubation method and the tracheotomy intubation method is relatively large.

[0071] Comparative studies have found that: This method greatly improves the efficiency of the tracheal intubation experiment of experimental mice and helps the progress of related experiments.

[0072] Taking the control experiment as a comparative example, if the reaction temperature is 20~50°C, test schemes corresponding to reaction temperatures outside the range of 20~50 can be provided.

[0073]

[0074] Note: Compared with the blind intubation method, ****P<0.01; compared with the transillumination intubation method through the neck, ****P<0.01, compared with the tracheotomy intubation method, ****P<0.01.

[0075] Regarding the selection of materials required for the above experiments, the following are used: 1. Blind intubation method: Stainless steel tracheal intubation for experimental mice: RWD; Toothless forceps: RWD.

[0076] 2. Transillumination intubation method through the neck: Forceps: RWD; Pliers: RWD; Light source: LED strong light source, RWD.

[0077] 3. Tracheotomy intubation method: Surgical suture No. 6, surgical needle No. 6: RWD; Ophthalmic scissors: RWD; 12 cm hemostatic forceps: RWD; Glass microdissection needle: RWD; 4. This method: Endoscope: 2.8mm customized endoscope; Anesthesia and respiration integrated machine: RWD; Robot arm: Customized DOFBOT robot arm (main board: Raspberry Pi 4B, computing power: 0.2 TFLOPS, CPU: Quad-Core Arm Cortex-A72 1.5GHz, GPU: Broadcom VideoCore IV, OLED device); AI chip: STM32 chip; Anesthesia maintenance box: Customized RWD small anesthesia gas release climate box with a volume of 0.128 cubic meters; Anesthetic gas: Isoflurane gas; Tracheal intubation: Customized RWD balloon tracheal intubation cannula (including guide wire, balloon, intubation).

[0078] In summary, the intelligent tracheal intubation surgery system and method for experimental mice based on robot arm linkage provided by the present invention use machine vision combined with deep learning CNN (i.e., Convolutional Neural Network) to dynamically identify and label the position of the tracheal duct in the laryngeal cavity of large / small / naked mice and the opening time of the tracheal duct, and at the same time link the robot arm to realize the dynamic insertion of tracheal intubation, realizing a fully automated tracheal intubation technology. Through artificial intelligence (AI) visual assistance, the experimental success rate is effectively improved, time is saved, and the risk of difficult exposure of the airway duct during intubation, failure of tracheal intubation of large / small / naked mice caused by improper operation of the operator, and various complications and even death due to the special anatomical characteristics of large / small / naked mice is prevented. In short, the present invention is a technology for identifying narrow cavities and assisting in automated tracheal intubation, which provides a more reliable and effective experimental scheme for researchers and a new idea for the development and progress of future animal experiments.

[0079] For the remaining technical features in this embodiment, those skilled in the art can flexibly select them according to the actual situation to meet different specific actual needs. However, it is obvious to those of ordinary skill in the art that these specific details do not have to be adopted to implement the present invention. In other instances, in order to avoid confusing the present invention, the well-known components, structures or parts are not specifically described, and they are all within the scope of the technical protection defined by the technical solution claimed in the claims of the present invention.

[0080] Modifications and variations made by persons skilled in the art without departing from the spirit and scope of the present invention shall fall within the scope of protection of the appended claims of the present invention. In the above description, in order to provide a thorough understanding of the present invention, a large number of specific details are set forth. However, it will be apparent to those of ordinary skill in the art that the present invention may be practiced without these specific details. In other instances, well-known technologies such as specific construction details, operating conditions and other technical conditions are not specifically described in order to avoid obscuring the present invention.

[0081] Specific examples are used herein to illustrate the principles and embodiments of the present invention. The description of the above embodiments is only for helping to understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific embodiments and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. An intelligent tracheal intubation surgical system for experimental mice based on robotic arm linkage, characterized in that: include: A respiratory anesthesia machine, an anesthesia maintenance box, and a mechanical arm. The anesthesia maintenance box is connected to the respiratory anesthesia machine through a rigid anesthetic gas communication pipeline. A mechanical arm moving track is arranged on the upper part of the anesthesia maintenance box. A retractable experimental mouse device plate for fixing the limbs of the experimental mouse is arranged in the anesthesia maintenance box. The operating port of the robot arm is provided with a tracheal tube release device, which is used to clamp the tracheal tube and move it deeper into the laryngeal cavity of the experimental mouse after the oral cavity of the experimental mouse is pried open; An image acquisition module is arranged beside the tracheal tube, and an image processing module is arranged in the mechanical arm; the image acquisition module is used to acquire continuous images of the laboratory mouse in vitro and in the oral cavity in real time, and transmit them to the image processing module; The image processing module is used to automatically analyze the continuous images in the oral cavity according to the image data, identify and mark the trachea of ​​the experimental mouse, and then automatically calculate the time difference between the opening and closing of the tracheal tube window of the experimental mouse; the robotic arm adjusts the position of the tracheal tube according to the position of the trachea of ​​the experimental mouse, automatically pauses the movement after determining the optimal tracheal intubation position, and uses the tracheal tube release device to insert the tracheal tube into the trachea of ​​the experimental mouse.

2. The intelligent tracheal intubation surgical system for experimental mice based on mechanical arm linkage according to claim 1 is characterized in that: The image processing module includes a preprocessing unit, a convolutional neural network unit, and a Python algorithm unit; the Python algorithm unit is used to calculate the time difference between the opening and closing of the tracheal tube window of the experimental mouse; the convolutional neural network unit includes an input layer, a convolution layer, a pooling layer, a fully connected layer, and an output layer; The preprocessing unit preprocesses the continuous images in the oral cavity to obtain a picture pixel matrix; the input layer is used to receive three consecutive frames of continuous images and extract five channel information of grayscale, horizontal coordinate gradient, vertical coordinate gradient, x-optical flow, and y-optical flow of each frame; the convolution layer is used to identify the spatial pattern in the picture pixel matrix and perform convolution operations on the 3D convolution kernels of the five channels to obtain features; the pooling layer is used to reduce the size of the matrix to reduce the parameters in the entire neural network; the fully connected layer is used to classify the features, obtain the classification results and transmit them to the output layer; the output layer is used to output the classification results in S4.4 to realize the identification and labeling of the trachea.

3. The intelligent tracheal intubation surgical system for experimental mice based on mechanical arm linkage according to claim 1 is characterized in that: The robotic arm is a multi-axis robotic arm; the robotic arm moving track includes a transverse track and a longitudinal track; the image acquisition module is a lens, and the lens is provided with an active lighting device; the tracheal intubation is an air bag tracheal intubation.

4. The intelligent tracheal intubation surgical system for experimental mice based on mechanical arm linkage according to claim 1 is characterized in that: The tracheal tube releasing device is a scissor-type tracheal tube releasing device, a tiny spring is placed in the center of the scissor-type tracheal tube releasing device, the scissor-type tracheal tube releasing device holds the tracheal tube so that it is located in the center of the tracheal tube releasing device and is connected to the spring, and the tracheal tube connected to the scissor-type tracheal tube releasing device tightly wraps the balloon trachea; the tracheal tube is a symmetrical hard half-cuff.

5. A method for intelligent tracheal intubation of experimental mice based on robot arm linkage, applied to the intelligent tracheal intubation system for experimental mice based on robot arm linkage as described in any one of claims 1 to 4, characterized in that: The following steps are involved: S1, turn on the respiratory anesthesia machine, which continuously injects isoflurane gas into the anesthesia maintenance box to ensure the anesthesia state of the experimental mouse until the muscle reflex of the experimental mouse disappears; S2, fix the limbs of the anesthetized experimental mouse to the retractable experimental mouse device board; S3, start the intubation operation: collect the in vitro continuous images of the experimental mouse through the image acquisition module, judge whether the tracheal tube release device reaches the oral cavity of the experimental mouse according to the in vitro continuous images, pry open the oral cavity of the experimental mouse with the tracheal tube release device to expose the larynx completely, and then clamp the tracheal tube and move it into the laryngeal cavity of the experimental mouse; S4, after being inserted into the oral cavity, the image acquisition module continues to collect continuous images of the oral cavity of the experimental mouse, and then the continuous images of the oral cavity are preprocessed by the image processing module, and the trachea of ​​the experimental mouse is identified and labeled based on the convolutional neural network, and the time difference of the opening and closing of the tracheal tube window of the experimental mouse is calculated; S5, the robotic arm adjusts the position of the tracheal tube according to the position of the trachea of ​​the experimental mouse, automatically stops moving after determining the optimal tracheal tube position, and inserts the tracheal tube into the mouse trachea using the tracheal tube release device according to the time difference; S6, the instruments except the endotracheal tube in the oral cavity of the experimental mouse were removed by a robotic arm, and the experimental mouse was manually taken out of the anesthesia maintenance box. After that, the respiratory anesthesia machine automatically stopped releasing isoflurane gas, completing the entire process of endotracheal intubation of the experimental mouse.

6. The intelligent tracheal intubation surgical method for experimental mice based on mechanical arm linkage according to claim 5 is characterized in that: In S4, the recognition and labeling of the trachea of ​​the experimental mouse is realized based on a convolutional neural network, which specifically includes the following steps: S4.

1. Preprocessing the continuous images of the oral cavity by a preprocessing unit in the image processing module to obtain a picture pixel matrix; S4.2, the input layer receives three consecutive frames of continuous images and extracts five channel information of grayscale, horizontal coordinate gradient, vertical coordinate gradient, x-ray flow, and y-ray flow of each frame; S4.3, identifying the spatial pattern in the pixel matrix of the image by the convolution layer, performing convolution operation on the 3D convolution kernels of the five channels, and obtaining the target features by extracting the image pixel matrix information; S4.4, reducing the size of the image pixel matrix by the pooling layer to reduce the parameters in the entire neural network; repeating steps S4.3 and S4.4; S4.5, the fully connected layer classifies the features, obtains the classification results and transmits them to the output layer; S4.

6. Finally, the output layer outputs the classification results in S4.4 to realize the identification and labeling of the trachea.

7. The intelligent tracheal intubation surgical method for experimental mice based on mechanical arm linkage according to claim 5 is characterized in that: In S4, calculating the time difference of opening and closing of the tracheal tube window of the experimental mouse specifically includes: After image recognition, the preprocessing unit extracts the image of the opening and closing of the airway window and transmits it to the Python algorithm unit in the image processing module. The Python algorithm unit automatically calculates the time difference of the opening and closing of the experimental mouse trachea based on the time difference between the transmitted images.

8. The intelligent tracheal intubation surgical method for experimental mice based on mechanical arm linkage according to claim 5 is characterized in that: The motion principle of the robotic arm is: The recorded image is grayed out, the center position of the robot's movement is calculated, and a three-dimensional coordinate axis is established based on the center position. The three-dimensional coordinate axis is used for trajectory planning and motion planning. The coordinates [x, y, z] are used to represent the spatial position points within the movable range of the robot arm to ensure that the operating ports of the robot arm can reach every spatial position point in the anesthesia maintenance box. A visual vision device is used to ensure the precise positioning and visual tracking of the experimental mouse by the robot arm.

Images