Gastroscope robot system based on visual touch fusion and optical learning

Abstract

The invention discloses a gastroscope robot system based on visual touch fusion and optical learning, and relates to the technical field of medical equipment.The gastroscope robot system comprises a gastroscope control system, a recognition and positioning module and an integral calculation module, the gastroscope control system controls and collects gastroscope image data, and the recognition and positioning module receives the gastroscope image data and sends the gastroscope image data to the integral calculation module; performing region-of-interest positioning on the gastroscope image data; and the integral calculation module performs molecular spectral measurement on the region of interest to obtain spectral data of stomach tissues, and performs integral energy ratio calculation based on the spectral data. High-efficiency endoscope motion control can be achieved, the burden of an operator is reduced, an advanced sensor and an execution mechanism are adopted, and high reliability and stability are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical equipment, and more specifically to a gastroscope robot system based on visual-tactile fusion and optical learning. Background Art

[0002] Upper gastrointestinal diseases are a widespread health challenge worldwide, especially gastric cancer, which ranks fifth and third in global cancer incidence and mortality, respectively. Gastroscopy is considered the gold standard for diagnosing and screening a variety of gastric diseases.

[0003] When doing a gastroscopy, the operator is often required to hold the insertion part of the gastroscope with one hand and the operating part of the gastroscope with the other hand, and keep the position of the gastroscope in the digestive tract and the lens angle fixed for a long time. Holding for a long time / maintaining an uncomfortable posture can easily cause fatigue to the operator, thereby affecting the accuracy of the examination and leading to missed diagnosis and misdiagnosis. At the same time, in conventional diagnosis and treatment, doctors mainly rely on visual feedback, especially in the context of endoscopic treatment. The operator's lack of experience leads to misjudgment, and hand shaking and displacement lead to positioning errors, which can easily cause iatrogenic injuries to patients. Therefore, it is particularly important to replace the doctor's two-handed operation of the gastroscope and high-precision positioning recognition based on artificial intelligence. Relevant patents have proposed gastroscope stents, which have freed the doctor's hands to a certain extent, but the stents on the market still have defects. For example, the doctor still needs to directly hold the operating part of the gastroscope to adjust the position and angle of the gastroscope, which is cumbersome and inefficient.

[0004] Therefore, how to replace the doctor's two-handed operation of the gastroscope and perform positioning and recognition of the image is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a gastroscopy robot system based on visual-tactile fusion and optical learning to solve the problems existing in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A gastroscope robot system based on visual-tactile fusion and optical learning includes: a gastroscope control system, an identification and positioning module, and an integral calculation module. The gastroscope control system controls the collection of gastroscope image data, the identification and positioning module receives the gastroscope image data, and locates the region of interest of the gastroscope image data; the integral calculation module performs molecular spectroscopy measurement on the region of interest to obtain spectral data of gastric tissue, and calculates the integral energy ratio based on the spectral data.

[0008] Preferably, the gastroscope control system includes: a gastroscope, a robotic arm, a micro force sensor, a rotor manipulator and a host, the gastroscope is arranged at the end of the robotic arm and is linearly connected to the host to collect gastroscope image data; the robotic arm is linearly connected to the host, and the movement of the robotic arm is controlled by a preset algorithm in the host; the micro force sensor is arranged at the end of the gastroscope, and when the end of the gastroscope contacts and squeezes the inner wall of the gastrointestinal tract, a force signal is generated and sent to the host; the rotor manipulator controls the angle of the gastroscope image.

[0009] Preferably, the gastroscope comprises: a gastroscope operating part, a gastroscope insertion part, an optical guide lens, an objective lens, and an image sensor. The gastroscope insertion part is a tubular structure equipped with an optical fiber. The gastroscope operating part is fixedly connected to the gastroscope insertion part. An equipment inlet is provided at the connection between the gastroscope operating part and the gastroscope insertion part. A major trochanter and a minor trochanter are provided at the other end of the gastroscope operating part. The major trochanter and the minor trochanter are coaxially arranged, and the major trochanter and the minor trochanter are connected to the guide wire of the gastroscope insertion part; a curved part is provided near the end of the gastroscope insertion part, and a guide wire is provided inside the gastroscope insertion part. The curved part at the end of the gastroscope insertion part is driven by twisting the guide wire by the major trochanter and the minor trochanter. The optical guide lens, the equipment outlet, the objective lens and the image sensor are arranged at the end of the gastroscope insertion part. The optical guide lens transmits the light emitted by the light source, the objective lens receives the light signal reflected by the gastrointestinal wall and forms an image, and the gastroscope image data is transmitted to the host through the image sensor; the endoscopic treatment instrument enters from the equipment inlet and extends from the equipment outlet.

[0010] Preferably, the robotic arm comprises: a position sensor, a force sensor, a motor, a base, two connecting rods, three joints, a clamp and a host connection port, the base is fixed with a joint, the connecting rods and the joints are spaced apart, and the clamp is arranged at the end of the joint farthest from the base; a motor, a position sensor and a force sensor are arranged at each joint, the motor controls the torsion of the joint, the position sensor obtains the motor rotation angle in real time, and the force sensor obtains the motor torque in real time; the host connection port is connected to the host through a line.

[0011] Preferably, the host comprises: a light source, a display screen, an operating platform, an image processor, a force processor, a visual-tactile fusion system and a gastroscope motion control system, wherein the light source transmits light signals to the light-guiding lens; the display screen displays image signals output by the image processor; the image processor receives gastroscope image data collected by the image sensor; the operating platform sends signals to the gastroscope motion control system; the force processor receives force signals collected by the micro force sensor, and establishes a safety and space model according to a preset tactile database; the visual-tactile fusion system fuses the gastroscope image data with the force signals, tracks the position and posture of the endoscope in the stomach in real time, and determines the pressure of the objective lens on the gastrointestinal wall; the gastroscope motion control system receives signals from the operating platform, converts the signals into motion commands through a preset control algorithm, and sends them to the rotor manipulator and the robotic arm.

[0012] Preferably, the rotor operator is provided with a recess matching the large rotor and the small rotor, comprising: an outer turntable, an inner turntable, a first servo motor, a second servo motor, a central axis and an outer shell, wherein the outer shell, the outer turntable and the inner turntable are coaxially arranged, and the central axis is fixed inside the outer shell; the first servo motor is arranged inside the outer turntable, and the second servo motor is arranged inside the inner turntable; the first servo motor and the second servo motor are both engaged with the central axis gear and rotate coaxially; the outer turntable drives the large rotor to rotate, and the inner turntable drives the small rotor to rotate.

[0013] Preferably, the rotor operator is cylindrical.

[0014] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a gastroscopic robot system based on visual-touch fusion and optical learning. The system can achieve efficient endoscope motion control and reduce the burden on the operator; automatically complete some tedious operations, saving the doctor's time and energy; adopt advanced sensors and actuators, with high reliability and stability; pave the way for the next step of remote laparoscope operation, and can achieve motion control according to network remote instructions. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0016] Figure 1 A schematic diagram of the structure provided by the present invention;

[0017] Figure 2Schematic diagram of the rotor operator provided by the present invention.

[0018] In the figure: 1 gastroscope; 2 robotic arm; 3 rotor manipulator; 4 miniature force sensor; 5 host; 11 gastroscope operating part; 12 gastroscope insertion part; 13 greater rotor; 14 smaller rotor; 15 bending part; 16 equipment entrance; 17 equipment exit; 18 optical guide lens; 19 objective lens; 191 image sensor; 21 base; 22 joint; 23 connecting rod; 24 clamp; 31 outer turntable; 32 inner turntable; 331 first servo motor; 332 second servo motor; 34 central axis; 35 outer shell; 51 light source; 52 display screen; 53 operating platform; 54 image processor; 55 force processor. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] An embodiment of the present invention discloses a gastroscope robot system based on vision-touch fusion and optical learning, including: a gastroscope control system, an identification and positioning module, and an integral calculation module. The gastroscope control system controls the collection of gastroscope image data, the identification and positioning module receives the gastroscope image data, and locates the region of interest of the gastroscope image data; the integral calculation module performs molecular spectral measurement on the region of interest to obtain spectral data of gastric tissue, and calculates the integral energy ratio based on the spectral data.

[0021] In a specific embodiment, Figure 1 As shown, the gastroscope control system includes: a gastroscope 1, a robotic arm 2, a rotor manipulator 3, a micro force sensor 4 and a host 5. The gastroscope 1 is arranged at the end of the robotic arm 2 and is linearly connected to the host 5 to collect gastroscope image data; the robotic arm 2 is linearly connected to the host 5, and the movement of the robotic arm 2 is controlled by a preset algorithm in the host 5; the rotor manipulator 3 is installed on the large and small rotors (detachable) to control the angle of the gastroscope image; the micro force sensor 4 is arranged at the end of the gastroscope 1, and when the end of the gastroscope 1 contacts and squeezes the inner wall of the gastrointestinal tract, a force signal is generated and sent to the host 5.

[0022] In a specific embodiment, the gastroscope 1 includes: a gastroscope operating part 11, a gastroscope insertion part 12, an optical guide lens 18, an objective lens 19, and an image sensor 191. The gastroscope insertion part 12 is a tubular structure equipped with an optical fiber. The gastroscope operating part 11 is fixedly connected to the gastroscope insertion part 12. An equipment entrance 16 is provided at the connection between the gastroscope operating part 11 and the gastroscope insertion part 12, which is used to penetrate into the patient's digestive tract, and endoscopic treatment instruments (such as forceps, etc.) enter through the entrance; a greater trochanter 13 and a smaller trochanter 14 are provided at the other end of the gastroscope operating part, and the greater trochanter 13 and the smaller trochanter 14 are coaxially arranged, and the greater trochanter 13 and the smaller trochanter 14 are connected to the guide wire of the gastroscope insertion part 12; a bending part 15 is provided near the end of the gastroscope insertion part 12, and a device entrance 16 is provided inside the gastroscope insertion part 12. There is a guide wire connected to the greater trochanter 13 and the lesser trochanter 14. The guide wire is twisted by the greater trochanter 13 and the lesser trochanter 14 to drive the curved portion 15 at the end of the gastroscope insertion portion 12, so as to adjust the rotation direction of the gastroscope lens. The optical guide lens 18, the equipment outlet 17, the objective lens 19 and the image sensor 191 are arranged at the end of the gastroscope insertion portion 12. The optical guide lens 18 transmits the light emitted by the light source 51. The optical guide lens 18 is the end of the optical fiber, and the light emitted by the transmitted light source illuminates the inside of the gastrointestinal tract. The objective lens 19 receives the light signal reflected by the gastrointestinal tract wall and forms an image. The gastroscope image data is transmitted to the host 5 through the image sensor 191. The endoscopic treatment instrument enters from the equipment inlet 16, extends from the equipment outlet 17, (such as grasping forceps, etc.) enters the gastrointestinal tract through the outlet to perform operations such as biopsy.

[0023] In a specific embodiment, the robotic arm 2 includes: a position sensor, a force sensor, a motor, a base 21, two connecting rods 23, three joints 22, a clamp 24 and a host connection port. The base 21 is fixed with a joint 22, the connecting rod 23 and the joint 22 are spaced apart, and the clamp 24 is arranged at the end of the joint 22 farthest from the base 21. The clamp 24 can be expressed as a handheld robotic arm structure or a splint-like structure. Regardless of the structural style, the clamp 24 is in contact with the gastroscope operating part 11, and is responsible for supporting and fixing the gastroscope 1, so as to facilitate cooperation with the robotic arm. 2's movement and posture change; each joint 22 is provided with a motor (so that the gastroscope 1 can complete forward, backward, external rotation, internal rotation and other movements), a position sensor and a force sensor (installed on the connecting shaft of the motor, used to obtain the rotation angle and torque of the motor in real time, and the formed position signal and force signal are transmitted to the host 5), the motor controls the torsion of the joint 22, the position sensor obtains the motor rotation angle in real time, and the force sensor obtains the motor torque in real time; the host connection port is connected to the host 5 through a line, which is responsible for the transmission of the position signal and the force signal, and the input of the motion control command.

[0024] In a specific embodiment, the host 5 includes: a light source 51, a display screen 52, an operating platform 53, an image processor 54, a force processor 55, a visual-tactile fusion system and a gastroscope motion control system. The light source 51 is a cold light source, which is responsible for outputting optical signals from the optical fiber to the optical lens 18; the display screen 52 displays the image signal output by the image processor 54, and displays it in the form of images / videos, etc.; the image processor 54 receives and processes the gastroscope image data input by the image sensor 191, and can decompose the position signal, part signal, lesion signal and posture signal according to the image data, form an identification and positioning model, and form a visualized image signal output to the display screen 52. At the same time, the image processor 54 includes a software part, that is, a high-precision tissue recognition and positioning algorithm based on artificial intelligence (see below for details), which can be converted into position signals, part signals, lesion signals and posture signals. etc., forming an identification and positioning model, and forming a three-dimensional image to be displayed on the display screen 52; the operation platform 53 is the doctor's control platform, and can input movement instructions such as forward / backward / ..., so as to realize remote, zero-contact, and accurate gastroscope operation, and send signals to the gastroscope motion control system; the force processor 55 receives the force signal collected by the micro-force sensor 4, and establishes a safety and space model according to the preset tactile database (including safety thresholds, tactile spatial features at different positions); the visual tactile fusion system is the software part, which integrates the features of different properties of visual tactile signals in stages, calculates the correlation between visual features and tactile, and realizes high-quality fusion of visual and tactile features with the maximum correlation coefficient as an indicator, which can realize real-time tracking of the position and posture of the endoscope in the stomach, and judge the pressure of the objective lens on the gastrointestinal wall, which can be displayed on the display screen 52, and provide early warning and navigation. The gastroscope motion control system receives the signal from the operation platform 53, converts the signal into a motion instruction through a preset control algorithm, and sends it to the rotor operator 3 and the robot arm 2. The gastroscope motion control system is the software part, which is mainly the control algorithm, including a two-layer closed-loop control structure: the inner layer is the closed-loop control of the desired posture and actual posture feedback of the lens, and the outer layer is the control of the motor output torque and the connecting shaft feedback torque.

[0025] In a specific embodiment, Figure 2As shown, the rotor operator 3 is provided with a recess matching the large rotor 13 and the small rotor 14, including: an outer turntable 31, an inner turntable 32, a first servo motor 331, a second servo motor 332, a central axis 34 and a housing 35. The housing 35, the outer turntable 31 and the inner turntable 32 are coaxially arranged, and the central axis 34 is fixed inside the housing 35; the first servo motor 331 is arranged inside the outer turntable 31, and the second servo motor 332 is arranged inside the inner turntable 32; the first servo motor 331 and the second servo motor 332 are both gear-engaged with the central axis 34 to rotate coaxially; the outer turntable 31 drives the large rotor 13 to rotate, and the inner turntable 32 drives the small rotor 14 to rotate. The rotation of the outer turntable 31 and the inner turntable 32 does not interfere with each other. The first servo motor 331 and the second servo motor 332 can receive the motion signal provided by the host 5 and generate power to drive the outer turntable 31 to rotate counterclockwise / clockwise and drive the inner turntable 32 to rotate counterclockwise / clockwise.

[0026] In a specific embodiment, the rotor operator 3 is cylindrical.

[0027] The micro force sensor 4 is a tiny device additionally installed on the gastroscope 1. It can be placed at the end of the gastroscope insertion part 12 in various forms such as pasting / magnetic attraction / clamping. When the gastroscope lens contacts and squeezes the inner wall of the gastrointestinal tract, a force signal is dynamically generated and transmitted back to the host 5 via the internal Bluetooth wireless.

[0028] The high-precision tissue recognition and positioning algorithm based on artificial intelligence can use the continuous dynamic image data obtained by gastroscopy to perform data processing and feature extraction, and combine it with the existing endoscopic image recognition and positioning algorithm, training and improvement, so that it can accurately identify and locate lesion features in a mixed background. The algorithm output can provide accurate target position information to the host 5, laying a foundation for accurate positioning of the gastroscopy robot system based on visual-touch fusion and optical learning. It mainly includes three parts.

[0029] In the identification and positioning module: Establish an endoscope image and video data database

[0030] ① Data acquisition and preprocessing: Use endoscopic equipment to obtain high-quality image and video data, and ensure standardization and quality control during data acquisition, such as calibrating equipment and light sources. Each data sample is labeled in detail, including information such as lesion location, size, severity, etc., and the accuracy and clinical relevance of the annotations are ensured through cooperation with clinicians. Because the data may have problems such as low contrast, illumination changes, motion blur, and noise, the data needs to be preprocessed to simplify the subsequent analysis process and improve the robustness of the algorithm.

[0031] ② Feature extraction: There are many important feature information in the pre-processed endoscopic images and video data. These features can include the tissue structure, morphological characteristics, dynamic changes and motion characteristics of the lesion, which are the basis of subsequent recognition and positioning algorithms and can help identify the area of ​​interest and accurately locate it.

[0032] ③ Database design and management: After data processing is completed, it is necessary to establish a suitable database to save and manage data to ensure effective support for the storage, management and analysis needs of medical data, while ensuring data security and privacy protection and complying with relevant laws and regulations and standards.

[0033] Establishing an endoscopic image and video data database

[0034] ① Algorithm selection and optimization: Existing recognition and positioning models are developing rapidly. Based on previous research, we evaluate the accuracy, recall rate, and processing speed of the models to provide useful guidance and reference for model selection. We select models suitable for different scenarios for further analysis.

[0035] ②Algorithm tuning and expansion: In view of the unique features and challenges of the gastric environment, such as lesion morphology, illumination changes and image noise under different viewing angles, the algorithms for different scenarios are fine-tuned and expanded. This includes optimizing the model architecture, adjusting the network depth and parameter settings, etc., to improve the accuracy and stability of recognition and positioning.

[0036] 3. Migration from Algorithms to Control System Integration

[0037] ① Real-time processing system: Endoscopic examination requires real-time feedback of positioning and identification information. It is necessary to comprehensively consider hardware acceleration, efficient algorithms, streaming processing architecture and user interface design. A high-performance real-time processing system suitable for endoscopic examination can be developed to ensure that medical personnel can quickly obtain and analyze key identification and positioning information.

[0038] ② System integration: Integrate the verified organ recognition and positioning model and real-time processing system into the gastroscope robot system to ensure that the gastroscope robot has accurate positioning recognition technology and real-time processing capabilities.

[0039] In the integral calculation module: Based on the optical-Raman spectroscopy, precise diagnosis and treatment technology for gastric lesions is used. The method of combining the fiber Raman spectroscopy system and the integral energy ratio is used to quickly locate gastric lesions, perform molecular spectral measurements of gastric tissue, obtain spectral data of gastric tissue, and perform real-time analysis and non-invasive diagnosis. It includes Raman spectrometer, Raman light source, Raman fiber probe and Raman data processing equipment. Effective integration of different components ensures that internal tissue result images, contact mechanics data, and atlas data can be collected simultaneously during gastroscopy operation, ensuring the stability and accuracy of the operation, and achieving real-time collection, analysis and diagnosis. It mainly includes two parts.

[0040] 1. Establishing a Raman Atlas Dataset of Gastric Lesions:

[0041] ① Data collection: Collect more than 30,000 spectral data of normal gastric tissue and abnormal gastric tissue. Obtain characteristic Raman spectra of normal gastric tissue and abnormal gastric tissue. In order to solve the problem of interference of Raman spectrum fluorescence background on the detection results, use a variety of data analysis methods to analyze and process the obtained tissue spectrum data, including denoising, feature extraction, statistical analysis, classification, prediction, etc.

[0042] ② Establishing a Raman spectrum data set of gastric lesions: The original Raman spectrum was processed and converted using fast Fourier transform, and the Raman spectrum features were extracted to establish a Raman spectrum data set of gastric lesions. The collected Raman spectrum was processed and compared with the Raman spectrum data set to obtain the integral energy ratio of the normal and lesion spectra. The integral energy ratio was used as an indicator for reference by medical staff.

[0043] 2. Spectral measurement and data processing

[0044] Using the biopsy hole of the gastroscope, fiber Raman technology is used to measure the molecular spectrum of gastric tissue during gastroscopy to obtain the spectrum data of normal and diseased gastric tissue. Using cluster analysis, principal component analysis, support vector machine, neural network and other algorithms, a mathematical model for identifying specific types of spectral features is established to accurately diagnose gastric cancer lesions and biopsy surgical margins.

[0045] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0046] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A gastroscope robot system based on visual-tactile fusion and optical learning, characterized in that: include: A gastroscope control system, an identification and positioning module, and an integral calculation module. The gastroscope control system controls the collection of gastroscope image data. The identification and positioning module receives the gastroscope image data and locates the region of interest of the gastroscope image data. The integral calculation module performs molecular spectroscopy measurement on the region of interest to obtain spectral data of gastric tissue, and calculates the integral energy ratio based on the spectral data.

2. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 1, characterized in that: The gastroscope control system includes: a gastroscope, a robotic arm, a micro force sensor, a rotor operator and a host. The gastroscope is arranged at the end of the robotic arm and is linearly connected to the host to collect gastroscope image data; the robotic arm is linearly connected to the host, and the movement of the robotic arm is controlled by a preset algorithm in the host; the micro force sensor is arranged at the end of the gastroscope, and when the end of the gastroscope contacts and squeezes the inner wall of the gastrointestinal tract, a force signal is generated and sent to the host; the rotor operator controls the angle of the gastroscope image.

3. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 2, characterized in that: The gastroscope comprises: a gastroscope operating part, a gastroscope insertion part, an optical guide lens, an objective lens, and an image sensor. The gastroscope insertion part is a tubular structure equipped with an optical fiber. The gastroscope operating part is fixedly connected to the gastroscope insertion part. An equipment inlet is arranged at the connection between the gastroscope operating part and the gastroscope insertion part. A large trochanter and a small trochanter are arranged at the other end of the gastroscope operating part. The large trochanter and the small trochanter are coaxially arranged, and the large trochanter and the small trochanter are connected to the guide wire of the gastroscope insertion part; a curved part is arranged near the end of the gastroscope insertion part, and a guide wire is arranged inside the gastroscope insertion part. The curved part at the end of the gastroscope insertion part is driven by twisting the guide wire by the large trochanter and the small trochanter. The optical guide lens, the equipment outlet, the objective lens and the image sensor are arranged at the end of the gastroscope insertion part. The optical guide lens transmits the light emitted by the light source, the objective lens receives the light signal reflected by the gastrointestinal wall and forms an image, and the gastroscope image data is transmitted to the host through the image sensor; the endoscopic treatment instrument enters from the equipment inlet and extends from the equipment outlet.

4. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 3, characterized in that: The robotic arm includes: a position sensor, a force sensor, a motor, a base, two connecting rods, three joints, a clamp and a host connection port, the base is fixed with a joint, the connecting rods and the joints are spaced apart, and the clamp is arranged at the end of the joint farthest from the base; each joint is provided with a motor, a position sensor and a force sensor, the motor controls the torsion of the joint, the position sensor obtains the motor rotation angle in real time, and the force sensor obtains the motor torque in real time; the host connection port is connected to the host through a line.

5. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 4, characterized in that: The host includes: a light source, a display screen, an operating platform, an image processor, a force processor, a visual-tactile fusion system and a gastroscope motion control system. The light source transmits light signals to the light-guiding lens; the display screen displays image signals output by the image processor; the image processor receives gastroscope image data collected by the image sensor; the operating platform sends signals to the gastroscope motion control system; the force processor receives force signals collected by the micro force sensor, and establishes a safety and space model according to a preset tactile database; the visual-tactile fusion system fuses gastroscope image data and force signals, tracks the position and posture of the endoscope in the stomach in real time, and determines the pressure of the objective lens on the gastrointestinal wall; the gastroscope motion control system receives signals from the operating platform, converts the signals into motion commands through a preset control algorithm, and sends them to the rotor manipulator and the robotic arm.

6. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 3, characterized in that: The rotor operator is provided with a recess matching the large rotor and the small rotor, and includes: an outer turntable, an inner turntable, a first servo motor, a second servo motor, a central axis and an outer shell. The outer shell, the outer turntable and the inner turntable are coaxially arranged, and the central axis is fixed inside the outer shell; the first servo motor is arranged inside the outer turntable, and the second servo motor is arranged inside the inner turntable; the first servo motor and the second servo motor are both engaged with the central axis gear and rotate coaxially; the outer turntable drives the large rotor to rotate, and the inner turntable drives the small rotor to rotate.

7. A gastroscope robot system based on visual-tactile fusion and optical learning according to claim 6, characterized in that: The rotor operator is cylindrical.

Images