Automatic positioning method and system for optical tracking positioning instrument of surgical robot

Abstract

The invention relates to an automatic positioning method of an optical tracking position indicator of a surgical robot, which comprises the following steps: when an automatic positioning system is started, acquiring current position information of the optical tracking position indicator; obtaining an identifier identification result of the tracking identifier of the optical tracking locator; obtaining position adjustment information based on the identification recognition result and the current position information; acquiring height adjustment data, first angle adjustment data and second angle adjustment data based on the position adjustment information; and adjusting the height of the lifting arm based on the height adjustment data, adjusting the first angle of the horizontal rotating arm based on the first angle adjustment data, and adjusting the second angle of the pitching arm based on the second angle adjustment data, so that the tracking identifier is located at a preset identification position of the optical tracking locator.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical robots, and particularly relates to an automatic positioning method and system of an optical tracking positioner of a surgical robot. BACKGROUND

[0002] With the rapid development of medical technology, surgical robots are increasingly widely used in clinical surgery. As a core component in a surgical robot system, an optical tracking positioner can track the positions and postures of surgical instruments and patient anatomical structures in real time and accurately, and provides key support for precise operation of surgery, and plays an indispensable role in surgical navigation and minimally invasive intervention.

[0003] At present, common installation structures of optical tracking positioners of surgical robots on the market are usually composed of multiple mechanical arms. These mechanical arms are designed to use traditional manual adjustment or relatively simple motor driving methods to adjust the spatial positions and angles of the optical tracking positioners.

[0004] However, manual adjustment depends on the experience and proficiency of operators, and the positioning process is time-consuming and can easily affect positioning accuracy. SUMMARY

[0005] The present application provides an automatic positioning method and system of an optical tracking positioner of a surgical robot, which aims to more accurately adjust the swing position of the optical tracking positioner by combining a tracking mark and the current positions of the mechanical arms, and to reduce the dependence on operators and the time loss caused by insufficient operator experience.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] An automatic positioning method of an optical tracking positioner of a surgical robot is applied to an automatic positioning system, the automatic positioning system comprising a lifting arm, a passive arm rotationally connected to the lifting arm, a horizontal rotation arm rotationally connected to the passive arm, a pitch arm rotationally connected to the horizontal rotation arm, and an optical tracking positioner fixedly connected to the pitch arm, and the automatic positioning method comprises the following steps:

[0008] When the automatic positioning system is started, current position information of the optical tracking positioner is acquired;

[0009] An identification recognition result of a tracking mark of the optical tracking positioner is acquired;

[0010] Position adjustment information is obtained based on the identification recognition result and the current position information;

[0011] acquire height adjustment data, first angle adjustment data and second angle adjustment data based on the position adjustment information;

[0012] adjust the height of the lifting arm based on the height adjustment data, adjust the first angle of the horizontal rotating arm based on the first angle adjustment data, and adjust the second angle of the pitching arm based on the second angle adjustment data, so that the tracking mark is located at the preset recognition position of the optical tracking positioner.

[0013] Correspondingly, the application also provides an automatic positioning system of the automatic positioning method of the optical tracking positioner of the surgical robot.

[0014] a base;

[0015] a lifting arm, one end of which is fixedly arranged on the base, and used for adjusting the height of the optical tracking positioner;

[0016] a passive arm, one end of which is rotationally connected to the end of the lifting arm away from the base, and the rotation axis of the passive arm is parallel to the lifting arm;

[0017] a horizontal rotating arm, one end of which is rotationally connected to the end of the passive arm away from the lifting arm, and the rotation axis of the horizontal rotating arm is perpendicular to the lifting arm;

[0018] a pitching arm, one end of which is rotationally connected to the end of the horizontal rotating arm away from the passive arm, and the other end of which is fixedly connected to the optical tracking positioner, and the rotation axis of the pitching arm is perpendicular to the horizontal rotating arm.

[0019] The application has the following advantages:

[0020] 1. After the system is started, the current position information of the optical tracking positioner is automatically acquired, and the position adjustment is completed in combination with the mark recognition result, without the need for the operator to manually adjust the complex joints of the mechanical arm. This eliminates the dependence on the experience and proficiency of the operator, and even non-technical personnel can complete the operation, which can reduce the dependence of the operation on personnel and improve the ease of use.

[0021] 2. The application adopts the mode of "automatic calculation + active adjustment": height adjustment data and angle adjustment data are automatically generated through a preset algorithm, and the lifting arm, the horizontal rotating arm and the pitching arm are driven to move cooperatively, without the need for manual trial and error. Compared with the "long time-consuming" of traditional manual adjustment, the process can realize rapid positioning, greatly shorten the positioning time and improve the surgical preparation efficiency.

[0022] 3、The application generates accurate adjustment data through coordinate system transformation, kinematics solving and other technologies based on the identification result and the current position information, ensuring that the optical tracking positioner can stably place the tracking mark at the preset identification position. This automatic and digital adjustment method avoids the operation errors in manual adjustment, improves the positioning accuracy, and ensures the reliability of the operation as much as possible.

[0023] 4、The adjustment data of the lifting arm, horizontal rotating arm and pitching arm is generated based on unified position adjustment information, and the actions of each component are coordinated and consistent, which can accurately adapt to the spatial requirements of different surgical areas. Compared with the difficulty of joint coordination in traditional manual adjustment, the system can more flexibly cope with complex surgical area environment, and ensure that the optical tracking positioner always maintains the best tracking angle as much as possible. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0025] Figure 1 The overall structure of an automatic positioning system of an optical tracking positioner of a surgical robot is provided for the embodiments of the present application.

[0026] Figure 2 The overall flowchart of an automatic positioning method of an optical tracking positioner of a surgical robot is provided for the embodiments of the present application.

[0027] Figure 3 The overall control flowchart of an automatic positioning system of an optical tracking positioner of a surgical robot is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0029] With the rapid development of medical technology, surgical robots are increasingly widely used in clinical operations. As a core component in the surgical robot system, the optical tracking positioner can track the position and attitude of surgical instruments and patient anatomical structures in real time and accurately, provide key support for precise operation of the operation, and play an indispensable role in surgical navigation and minimally invasive intervention.

[0030] At present, the common installation structure of the optical tracking positioner of the surgical robot on the market is usually composed of multiple mechanical arms. In the design, the spatial position and angle of the optical tracking positioner are adjusted by traditional manual adjustment or relatively simple motor driving mode.

[0031] However, manual adjustment depends on the experience and proficiency of the operator, the positioning process is time-consuming, and the positioning accuracy is easily affected. Therefore, the embodiment of the present application provides an automatic positioning method and system for the optical tracking positioner of the surgical robot, which aims to more accurately adjust the swing position of the optical tracking positioner by combining the tracking mark and the current position of each mechanical arm, and reduce the dependence on the operator, thereby reducing the time loss caused by the lack of experience of the operator.

[0032] Reference Figure 1 The embodiment of the present application provides an automatic positioning system for the optical tracking positioner of the surgical robot, comprising:

[0033] a base 1;

[0034] a lifting arm 2 fixed at one end to the base 1, used for adjusting the height of the optical tracking positioner;

[0035] a passive arm 3 rotatably connected at one end to the end of the lifting arm 2 away from the base 1, the rotation axis of the passive arm 3 being parallel to the lifting arm 2;

[0036] a horizontal rotating arm 4 rotatably connected at one end to the end of the passive arm 3 away from the lifting arm 2, the rotation axis of the horizontal rotating arm 4 being perpendicular to the lifting arm 2;

[0037] a pitch arm 5 rotatably connected at one end to the end of the horizontal rotating arm 4 away from the passive arm 3, and fixedly connected at the other end to the optical tracking positioner, the rotation axis of the pitch arm 5 being perpendicular to the horizontal rotating arm 4.

[0038] Specifically, the automatic positioning system comprises the base 1, the lifting arm 2, the passive arm 3, the horizontal rotating arm 4, the pitch arm 5 and the optical tracking positioner. The lower surface of the base 1 is provided with a roller with a locking function, which facilitates the movement of the base 1 and also locks the roller through the locking function of the roller, so as to keep the base 1 stable.

[0039] The lifting arm 2 is vertically fixed to the upper surface of the base 1, and comprises a coaxially arranged upper arm and a lower arm, the upper arm and the lower arm are fixed through a lifting shaft 21, the lower arm is vertically fixed to the upper surface of the base 1, the lifting shaft 21 can be an electric shaft or a hydraulic shaft, and can be controlled by a servo motor or an electric cylinder, and the embodiment of the application does not limit this, the height of the upper arm can be changed through the lifting shaft 21, thereby the height of the lifting arm 2 can be adjusted, and finally the height of the terminal optical tracking positioner is adjusted.

[0040] The top end of the upper arm is rotationally connected with the passive arm 3 through a first passive shaft 33, the passive shaft is parallel to the lifting arm 2, and the passive arm 3 can be rotated around the lifting arm 2 through manual adjustment, so that the position of the optical tracking positioner can be coarsely adjusted.

[0041] Further, in an embodiment, the passive arm 3 comprises a first arm 31 and a second arm 32.

[0042] One end of the first arm 31 is rotationally connected with the lifting arm 2, and the other end is rotationally connected with the second arm 32, the end of the second arm 32 away from the first arm 31 is rotationally connected with the horizontal rotating arm 4, and the rotation axis of the first arm 31 is parallel to the rotation axis of the second arm 32.

[0043] The first arm 31 is rotationally connected with the upper arm through the first passive shaft 33, the first arm 31 is vertically arranged with the lifting arm 2, and the end of the first arm 31 away from the upper arm is rotationally connected with the second arm 32 through the second passive shaft 34, the first passive shaft 33 is parallel to the second passive shaft 34, that is, the rotation axis of the first arm 31 is parallel to the rotation axis of the second arm 32, and the first arm 31 and the second arm 32 are arranged at an angle, which can be set according to actual conditions, and the embodiment of the application does not limit this. By adjusting the horizontal positions of the first arm 31 and the second arm 32, the position of the optical tracking positioner can be better adjusted, so that it can better point to the tracking mark. It should be understood that the first arm 31 and the upper arm, and the first arm 31 and the second arm 32 are both provided with positioning mechanisms, which can position the first arm 31 and the second arm 32, and ensure the stability of the passive arm 3, and the positioning mechanism can select a conventional general mechanism, and the embodiment does not limit this.

[0044] One end of the horizontal rotating arm 4 is rotationally connected with the second arm 32 through a horizontal rotating shaft 41, the horizontal rotating shaft 41 is parallel with the first arm 31, that is, perpendicular to the lifting arm 2. One end of the pitching arm 5 is rotationally connected with the end of the horizontal rotating arm 4 away from the second arm 32 through a pitching shaft 51, the pitching shaft 51 is perpendicular to the lifting arm 2, and perpendicular to the horizontal rotating shaft 41. The other end of the pitching arm 5, that is, the terminal end, is fixedly connected with the optical tracking positioner 6, and the pitching angle of the optical tracking positioner 6 can be adjusted by adjusting the angle of the pitching shaft 51. The horizontal rotating shaft 41 and the pitching shaft 51 can both control the rotating angle through a private motor, so as to realize the adjustment of the angle and position of the optical tracking positioner 6.

[0045] Further, with reference to Figure 2 The embodiment of the application further provides an automatic positioning method of an optical tracking positioner of a surgical robot, applied to an automatic positioning system, and the automatic positioning method comprises the following steps.

[0046] 101. When the automatic positioning system is started, the current position information of the optical tracking positioner is acquired.

[0047] Specifically, after the system is started, the sensors distributed at key positions are triggered to work synchronously: the laser displacement sensor at the bottom of the lifting shaft collects the real-time height in the vertical direction (Z axis) (such as 1750 mm), the absolute encoders at the joints of the horizontal rotating shaft and the pitching shaft collect the current rotating angles (such as the horizontal rotating angle 0° and the pitching angle 2°), and the IMU (inertial measurement unit) built in the optical tracking positioner records the real-time tilt posture of the device (such as the yaw angle 1° around the Z axis). At the same time, the system synchronizes the data of all the sensors through a time stamp, so as to ensure that the collected height, angle and posture parameters correspond to the state at the same time, and avoid the deviation caused by data delay. For example, when the neurosurgical system is started, the laser displacement sensor feeds back Z=1750 mm, the encoders feed back the horizontal rotation 0° and the pitching 2°, and the IMU feeds back the yaw 1°, and these data form the original position data set after synchronization.

[0048] The collected raw data is uniformly converted to the initial coordinate system of the system (the base coordinate system of the robot arm): taking the center point of the base as the origin, the Z axis as the vertical direction, the X axis pointing to the surgical area, and the Y axis perpendicular to the X axis to form a right-handed coordinate system. The height of the Z axis of the laser displacement sensor is directly mapped to the Z coordinate in the coordinate system; the horizontal rotation angle and the pitch angle are converted to the projection offset of the X-Y plane through trigonometric functions (for example, a pitch of 2° corresponds to a small offset in the Y axis direction); the yaw angle of the IMU is used to correct the overall attitude. After fusion, the three-dimensional coordinates of the tracker in the initial coordinate system of the system (such as X=500mm, Y=300mm, Z=1750mm) and the complete attitude parameters (0° around the X axis, 2° around the Y axis, 1° around the Z axis) are generated. For example, if the horizontal rotation axis rotates by 30°, through coordinate conversion, it can be known that the tracker deviates by 250mm in the X axis direction and 433mm in the Y axis direction (based on the arm length of 1000mm trigonometric function calculation), and finally fused into the three-dimensional coordinates.

[0049] The system calls the pre-stored device calibration parameters (such as sensor zero drift, robot arm length error) to correct the fused position information: the ±1mm error of the laser displacement sensor caused by temperature drift is eliminated through linear compensation; the ±0.1° error caused by the mechanical gap of the encoder is corrected through the backlash compensation algorithm. At the same time, the optical tracking positioner calculates the deviation of its own position from the theoretical value by shooting the fixed markers pre-set on the operating room ceiling (whose absolute position in the initial coordinate system of the system is known), and if the deviation exceeds 0.5mm or 0.1°, a secondary calibration is triggered until the accuracy requirement is met. For example, after shooting the ceiling markers, it is found that the actual position of the X axis deviates from the theoretical value by 2mm, and the system automatically corrects the X coordinate from 500mm to 502mm, ensuring the accuracy of the initial position information.

[0050] The traditional scheme only collects data once when it starts, without considering sensor drift; the present application can real-time correct the errors caused by temperature and mechanical deformation (such as the zero point offset of the laser sensor in a high temperature environment) through the secondary verification of the fixed markers, ensuring that the initial position information remains reliable in complex surgical environments, and is more practical.

[0051] Further, in some embodiments, when the automatic positioning system is started, the current position information of the optical tracking positioner is obtained, including:

[0052] When the automatic positioning system is started, the current state of the active joint in the automatic positioning system is parameterized based on a pre-set kinematic model, to obtain a transformation relationship from the initial pose of the system to the end pose;

[0053] Based on a pre-established connection relationship model, the fixed relationship between the end and the optical tracking positioner is analyzed to obtain a transformation relationship from the end to the optical tracking positioner;

[0054] Based on the preset coordinate transformation theory, the transformation relationship from the initial pose of the system to the end pose and the transformation relationship from the end to the optical tracking position instrument are synthesized to obtain the current position information of the optical tracking position instrument in the initial coordinate system of the system.

[0055] Specifically, through the kinematic model of the mechanical arm, the real-time state of the active shaft (lifting shaft, horizontal rotation shaft, and pitch shaft) is converted into the spatial position relationship of the system initial coordinate system (mechanical arm base coordinate system) to the end effector. Specifically, the system will collect the current parameters of the three active shafts in real time: the height value of the lifting shaft (such as the current height of 1500mm), the rotation angle of the horizontal rotation shaft (such as the current clockwise rotation of 30°), and the pitch angle of the pitch shaft (such as the current pitch of 15°). Based on the preset improved DH parameter model (a standard model describing the position relationship between the joints of the mechanical arm), through forward kinematics calculation, the parameters of these shafts are converted into the three-dimensional coordinates (such as X=500mm, Y=300mm, Z=1500mm) and the attitude (such as rotation of 30° around the X axis and rotation of 15° around the Y axis) of the end effector in the initial coordinate system of the system, and finally the transformation relationship between the two (i.e., the position and attitude of the end effector relative to the base) is formed.

[0056] Since the end effector and the optical tracking position instrument (camera) are rigidly connected (such as fixed by a mechanical support), the relative position and attitude of the two will not change, and this relationship needs to be determined in advance through hand-eye calibration. The process of hand-eye calibration is: control the end effector to drive the camera to shoot the same set of markers at multiple different positions, and calculate the fixed offset of the camera relative to the end effector through an algorithm (such as the camera center located at X=50mm, Y=0mm, Z=100mm in the end effector coordinate system, and the attitude remains the same). The system directly calls the pre-stored fixed relationship to analyze the transformation relationship from the end effector to the camera (i.e., the position and attitude of the camera relative to the end effector).

[0057] In the coordinate system transformation, the end effector is the intermediate link connecting the mechanical arm base coordinate system and the camera coordinate system: the transformation relationship from the mechanical arm base coordinate system to the end effector can be calculated through the forward kinematics of the mechanical arm, and then the transformation relationship from the mechanical arm base coordinate system to the camera coordinate system can be deduced by combining the fixed transformation relationship of the end effector and the camera, thereby providing the core coordinate reference for the position calculation and attitude adjustment of the optical tracking position instrument.

[0058] Based on the superposition principle of coordinate transformation (i.e. the relationship of "base→end effector" and the relationship of "end effector→camera" are superimposed, and the relationship of "base→camera" can be obtained), the system synthesizes the two transformation relationships. For example, if the end effector is located at (500, 300, 1500) in the base coordinate system, and the offset of the camera relative to the end effector is (50, 0, 100), then the position of the camera in the base coordinate system is (500+50, 300+0, 1500+100) = (550, 300, 1600), and combined with the superposition calculation of the attitude, the complete position information of the camera in the initial coordinate system of the system is obtained, that is, the current position information.

[0059] The technical scheme provided by the embodiments of the present application has the following technical benefits:

[0060] 1. In the traditional scheme, the position of the optical tracking positioner usually depends on mechanical scales or manual markers, and the coordinate references of different components are not unified, which is easy to produce cumulative errors. However, the present application unifies all position information to the same reference (the base coordinate system of the robot arm) through the chain transformation of "initial coordinate system of the system→end effector→camera", ensuring the accuracy of the camera position calculation and avoiding the errors caused by multiple references.

[0061] 2. In the traditional scheme, the joint angle of the robot arm and the distance between the camera and the end effector need to be manually measured, which is not only low in efficiency, but also easy to introduce human reading errors. The present application automatically collects the parameters of the active shaft through sensors and directly calls the pre-calibrated fixed relationship, without human intervention throughout, greatly improving the efficiency and accuracy of data collection.

[0062] 3. In the traditional scheme, if the attitude of the robot arm changes, the camera position needs to be manually measured again, which cannot respond in real time. The present application can calculate the transformation relationship based on the current state of the active shaft in real time. As long as the attitude of the active shaft changes (such as adjusting the position during surgery), the system can immediately update the position information of the camera, ensuring the real-time positioning.

[0063] 4. In the traditional scheme, the operator needs to master professional knowledge such as robot kinematics and coordinate transformation to calculate the camera position, which has a high threshold. The present application encapsulates the complex calculation process as an automated process, and the technical personnel only need to start the system to obtain accurate position information without understanding the underlying principles, reducing the professional requirements for the operator.

[0064] Further, in some embodiments, before obtaining the current position information of the optical tracking positioner, the method further comprises:

[0065] obtaining the passive adjustment result of the passive arm;

[0066] obtaining the current position information of the optical tracking positioner based on the passive adjustment result.

[0067] Specifically, the passive arms are two passive axes with one degree of freedom in the system, and the physician roughly points the optical tracking positioner to the direction of the operation area by manually rotating the passive joints. The system detects the rotational angle changes of the passive axes in real time through angle sensors (such as encoders) installed at the joints of the passive arms. For example, the physician rotates the left passive arm clockwise by 20° and the right passive arm counterclockwise by 15°, and the sensors will immediately record these two angle values and convert them into digital signals to transmit to the control system as the original data of passive adjustment. These data directly reflect the rough pointing of the optical tracking positioner on the horizontal plane by the passive arms and are the basis for subsequent accurate position calculation.

[0068] Since the passive arms are in series with the active axes (the lifting axis, the horizontal rotation axis, and the pitch axis), the rotation of the passive arms will change the initial posture of the entire mechanical arm. For example, after the passive arms are rotated clockwise by 20°, the entire active axis system (including the lifting axis, the horizontal rotation axis, and the pitch axis) will be deflected by 20° on the horizontal plane as a whole with the passive arms. At this time, the system will add the passive axis angle values obtained above as the initial offset to the calculation of the current position: when calculating the "transformation relationship from the initial posture of the system to the end posture", the initial pointing of the mechanical arm base coordinate system is first corrected according to the rotational angle of the passive arms; when synthesizing the camera position information, the horizontal rotation amount caused by the passive adjustment is taken into account in the coordinate transformation, and finally the current position of the optical tracking positioner containing the passive adjustment is obtained. For example, if the position of the camera in the initial coordinate system of the system is (550, 300, 1600) when the passive arms are not adjusted, after the passive arms are rotated clockwise by 20°, the system will correct this position to (550×cos20°-300×sin20°, 550×sin20°+300×cos20°, 1600) through coordinate rotation algorithm, so as to ensure that the position information is consistent with the actual posture after passive adjustment.

[0069] In the embodiments of the present application, the passive arms quickly point the tracking instrument to the general direction of the operation area, and the system takes the influence of passive adjustment into account in the position calculation through the coordinate correction mechanism, so that the reference for subsequent active fine adjustment is consistent with the actual posture after passive rough adjustment.

[0070] Further, in some embodiments, when the automatic positioning system is started, it further includes:

[0071] In response to the automatic search instruction, the identification recognition data of the optical tracking positioner in the field of view is obtained in real time based on the visual detection module;

[0072] Based on the current position information of the optical tracking positioner in the initial coordinate system of the system, the current posture parameters of the optical tracking positioner are calculated in combination with the forward kinematics model;

[0073] Based on the current pose parameters and the preset search range constraint conditions, a search path covering the possible region of the operation area is planned, and the search path includes the moving track of the optical tracking positioner in the lifting shaft direction, the rotating track of the horizontal rotating shaft, and the rotating track of the pitching shaft.

[0074] According to the search path, the control module sends a motion instruction to each active shaft to drive the optical tracking positioner to perform a search motion, and simultaneously continuously detects whether the tracking marker appears in the field of view to obtain a marker recognition result.

[0075] Specifically, after the manual adjustment of the passive arm coarse positioning is completed, the system enters an automatic search state, and the binocular vision system of the optical tracking positioner is used to monitor whether a tracking marker exists in the field of view in real time. After the manual coarse positioning is completed (the doctor has pointed the tracking instrument to the general direction of the operation area through the passive arm), the system automatically activates the visual detection module, the binocular camera starts to continuously collect images of the operation area, and whether a target matching the preset marker shape and size exists in the image is analyzed through a preset marker recognition algorithm (such as feature point matching, contour detection). For example, in orthopedic surgery, the marker pasted on the operation area is a black circular pattern, and the visual detection module scans the image frame by frame. If a region matching the feature is detected, the coordinate position of the region in the image is recorded; if no marker is detected, a result of “no marker recognized” is returned, which provides a basis for subsequent path planning.

[0076] Then, the spatial position of the tracking instrument is converted into a pose description that can be used for path planning, and a “starting point coordinate” is provided for the search path. The system calls the current position information of the tracking instrument calculated above (such as three-dimensional coordinates in the initial coordinate system of the system: X=600 mm, Y=400 mm, Z=1700 mm), and combines a forward kinematics model (a model describing the relationship between the motion of each shaft and the end pose) to analyze the current pose of the tracking instrument, including the horizontal rotation angle (rotating 30° around the X axis), the pitching angle (rotating 20° around the Y axis), and the position in the height direction. For example, if the tracking instrument currently points to the left side of the waist of the patient, the forward kinematics calculation will convert it into the quantitative parameters of “horizontal rotation angle 30° (leftward deviation), pitching angle 20° (downward inclination), and height 1700 mm”. These parameters clearly define the spatial pointing direction of the tracking instrument and are the basis for planning the search path.

[0077] Then, the ordered search trajectory is generated within the safety range to ensure efficient coverage of the area where the marker may exist in the operation area. The system first reads the preset search range constraints (such as the maximum moving range of the lifting shaft 1500-2000 mm, the rotating range of the horizontal rotating shaft ±60°, and the rotating range of the pitching shaft ±45° to avoid collision with the patient or the instrument), and then plans the path according to the "spiral expansion" or "partition scanning" strategy starting from the current pose. For example, if the current tracker does not detect the marker, the system will plan: first control the lifting shaft to gradually rise from 1700 mm to 1900 mm (each time by 50 mm), while the horizontal rotating shaft is rotated from 30° to the right side to-30° (each time by 10°), and the pitching shaft remains unchanged at 20°; if it is still not detected, keep the height at 1900 mm, rotate the pitching shaft from 20° to 45° upward (each time by 5°), and repeat the ±30° scanning of the horizontal rotating shaft. Such path design not only ensures comprehensive coverage of the operation area (such as the patient's waist and surrounding area), but also avoids the mechanical arm from exceeding the safety range through the constraint conditions, realizing accurate and safe search.

[0078] Finally, the search path execution link is executed through closed-loop control to realize dynamic adjustment of "motion-detection-feedback". The control module (such as the motor controller based on CANopen protocol) decomposes the above planned trajectory into specific action instructions of each active shaft: for example, the lifting shaft rising by 50 mm instruction is converted into the rotation number of the servo motor, and the horizontal rotating shaft rotating by 10° instruction is converted into the angle parameter of the corresponding motor. When the motor drives the mechanical arm to move according to the instruction, the vision detection module synchronously collects images and analyzes in real time. If the marker is detected to appear in the image center when the lifting shaft moves to 1800 mm and the horizontal rotating shaft rotates to 0°, the search is immediately stopped, and the current pose is recorded as the "marker identification result"; if it is still not detected after completing the entire path, it returns to "no marker identified", prompting the doctor to check the marker position. For example, in neurosurgery, when the tracker scans to the position directly above the patient's head and detects the marker, the system will immediately stop moving to ensure that the subsequent fine positioning starts from this accurate position.

[0079] The technical scheme provided by the embodiment of the application has the following technical benefits:

[0080] 1. Traditional manual adjustment is easy to cause incomplete search range (miss marker) or exceed the safety range (collision with the patient) due to insufficient doctor experience; the constraint conditions and ordered path of the application ensure coverage of all possible areas in the operation area, while strictly limiting within the safety range, and the reliability is higher.

[0081] 2. In traditional methods, doctors need to constantly monitor the tracking device screen and manually adjust the settings, which is very time-consuming. This application automates the entire process from "automatic startup to automatic planning, automatic execution, and automatic detection," allowing doctors to intervene only when prompted by the system, thus reducing their workload.

[0082] 3. Traditional manual search requires manual recalibration of the position, which is prone to deviation; this application directly outputs the "identifier recognition result", which can be directly used as the input for the step "obtain position adjustment information based on the identifier recognition result", ensuring a smooth process from search to fine-tuning and reducing intermediate errors.

[0083] 4. Traditional methods rely on doctors to manually rotate the stent repeatedly to find the marker, which often takes too long due to random orientation and lack of experience. In contrast, this application uses a pre-set path for orderly scanning, combined with real-time feedback from visual detection, which can reduce unnecessary time waste and thus improve recognition efficiency.

[0084] 102. Obtain the identification results of the tracking markers of the optical tracking and positioning device.

[0085] This includes: based on marker feature templates, performing feature matching processing on surgical area images acquired by an optical tracking and positioning device to generate candidate regions for potential markers in the images;

[0086] Based on binocular vision depth detection technology, the candidate region is subjected to three-dimensional spatial verification processing to screen out the real landmark region that meets the preset size and spatial location range.

[0087] Based on the mapping rules between image coordinates and spatial coordinates, coordinate transformation is performed on the real marker area to obtain the position and attitude information of the tracking marker in the coordinate system of the optical tracking and positioning instrument, i.e., the marker recognition result.

[0088] Specifically, first, through the preset marker feature template, the region where the marker may exist is quickly locked from the image of the surgical area taken by the optical tracking locator (binocular camera). The marker feature template is the unique visual features of the tracking marker stored in advance, for example, if the marker is a square plate printed with a specific two-dimensional code pattern, the template will record the texture distribution, edge contour, corner point position, etc. of the two-dimensional code. After the system starts, the binocular camera will take a static image of the surgical area (such as the surgical area of the patient's head and the surrounding environment), and the software algorithm will scan the image pixel by pixel, and compare the visual features (such as edge shape, texture density) of each region with the template features. For example, when the edge of a certain region in the image presents a square contour, and the internal texture matches the black and white distribution pattern of the two-dimensional code template with a matching degree of more than 90%, the system will mark this region as a "potential marker candidate region". If there are multiple similar patterns in the surgical area (such as patterns on the surgical towel), the system will retain all regions that meet the matching degree, forming a candidate list to provide a basis for subsequent screening.

[0089] Then, through the three-dimensional perception ability of binocular vision, the interference items in the candidate region are removed, and only the real marker that meets the spatial characteristics of the surgical area is retained. The two lenses of the binocular camera have a certain distance, and when taking the same object, a parallax will be generated, and the system can calculate the actual distance (distance from the camera) and physical size (such as length, width) of the candidate region in three-dimensional space through the parallax. For example, the preset reasonable range of the surgical area is "500-1500 mm from the camera, and the physical size of the marker is 40 mm x 40 mm ± 5 mm", if the candidate region A is 2000 mm away from the camera in three-dimensional space (out of range), or the physical size of the candidate region B is 100 mm x 100 mm (not consistent with the preset), it is determined as an interference region (such as a distant instrument reflection, a surgical towel wrinkle) and is removed; only when the distance of the candidate region C is 800 mm (within the range) and the size is 42 mm x 41 mm (consistent with the error requirement), it is confirmed as a "real marker region". This way solves the misidentification problem of "visual similarity but spatial inconsistency" in the plane image through three-dimensional space parameter verification.

[0090] Then, the marker position in the image is converted into a quantitative parameter in the optical tracking positioning instrument coordinate system, to provide accurate data for subsequent positioning adjustment. The mapping rule of image coordinates and space coordinates is determined in advance through camera calibration (such as the conversion ratio of pixel coordinates and millimeter coordinates, and the distortion correction parameters of the camera lens). For example, the pixel coordinates of the real marker region in the left camera image are (320, 240), and the pixel coordinates of the real marker region in the right camera image are (300, 240). The system combines binocular disparity and the mapping rule to calculate the three-dimensional coordinates of the marker in the optical tracking positioning instrument coordinate system (with the camera optical center as the origin, the X-axis horizontally to the right, the Y-axis vertically upward, and the Z-axis pointing to the surgical area): X = 100 mm, Y = -50 mm, and Z = 800 mm (indicating 100 mm to the right of the camera, 50 mm below, and 800 mm in front). At the same time, the attitude parameters (such as a 10° rotation around the Z-axis) are determined by identifying the angle of the marker edge (such as the inclination direction of the two-dimensional code). Finally, these position and attitude information are integrated into the “marker identification result” and directly used for subsequent position adjustment calculation.

[0091] The technical scheme provided by the embodiment has the following technical benefits:

[0092] 1. The traditional scheme relies on single image feature recognition, which is easily affected by the environment of the surgical area (such as the reflection of the surgical lamp and the shielding of the instrument), leading to misjudgment. The present application uses “feature matching + three-dimensional verification” for double screening, first locks the candidate from the plane image, and then removes the interference through spatial parameters. Even if there are similar patterns in the surgical area, the real marker can be accurately identified, and the misidentification rate is greatly reduced.

[0093] 2. The traditional scheme can only determine whether the marker is in the field of view, and cannot provide specific position and attitude data. The present application outputs the three-dimensional position and attitude of the marker in the camera coordinate system through coordinate conversion, provides quantitative basis for subsequent “adjusting the marker to the preset position”, and ensures that the adjustment of the optical tracking positioning instrument is more accurate.

[0094] 3. The embodiment focuses on “accurate analysis of static images”, which is suitable for scenarios where the marker is already in the field of view but needs to be accurately quantified (such as quickly confirming the marker parameters after manual coarse adjustment). It can be combined with the above-mentioned “detection in motion” and is suitable for the search stage where the marker position is unknown. The two complement each other and improve the adaptability of the system in different scenarios.

[0095] 4. The traditional scheme is sensitive to light and angle (such as being easily identified when the marker is inclined). The three-dimensional verification of the present application is not affected by light (depending on disparity calculation rather than brightness), and the coordinate conversion can be compatible with different attitudes of the marker. Even if the marker is slightly inclined or partially blocked, accurate parameters can still be output, ensuring the stability of positioning during surgery.

[0096] Further, in some embodiments, after obtaining the identification recognition result of the tracking mark of the optical tracking positioner, further comprising:

[0097] If the identification recognition result is that the tracking mark is not detected, the detection operation on the tracking mark is re-executed for a preset number of times and then stopped;

[0098] If the identification recognition result is that the tracking mark is detected, position adjustment information is obtained based on the identification recognition result and the current position information.

[0099] Specifically, if the result is that the tracking mark is not detected, the system starts a retry mechanism. At this time, the system does not immediately terminate the detection, but first analyzes the possible reasons for not detecting (such as the marker being temporarily blocked, the image being blurred due to instantaneous changes in light), and optimizes the detection conditions accordingly. For example, automatically adjusting the exposure parameters of the optical tracking positioner (increasing the brightness to cope with dark environments), expanding the feature matching threshold (from 90% to 85% to accommodate slight occlusions), or controlling the passive arm to fine-tune 1-2° (driving the tracking instrument to rotate slightly to avoid temporary occlusions). After that, the system re-executes the complete identification recognition process and records the number of retries. The preset number of times can be flexibly set according to the surgical scene (such as 3 times for neurosurgery which requires high precision, and 5 times for orthopedic surgery), and the interval between each retry is 0.5 seconds (to avoid frequent operations affecting stability). For example, if the first detection fails due to temporary occlusion of the surgical instrument to the marker, the system retries after adjusting the exposure, and the second detection successfully identifies the marker, then the retry is stopped and the subsequent process is entered; if the marker is not detected for 3 consecutive retries (such as the marker falling off), the system will trigger an audible and visual alarm to prompt the doctor to check the marker status, and record the failure log (including the reason analysis for each undetected).

[0100] The preset number of times can be flexibly set according to the surgical scene. For example, in neurosurgery, the surgical area (such as the intracranial space) is narrow, the precision requirement is extremely high (millimeter level or even sub-millimeter level), and the marker is easily disturbed by factors such as brain tissue movement and instrument occlusion. Therefore, more retry times (such as 5-8 times) are preset to reduce the probability of accidental failure. At the same time, in order to avoid the impact of frequent retries on the stability of the surgical area (such as mechanical arm vibration disturbing the brain tissue), a longer interval time (such as 1-2 seconds) is set.

[0101] In orthopedic surgery, the surgical area (such as the limbs and spine) is relatively open, the precision requirement is slightly lower than that of neurosurgery (millimeter level), but the marker may be disturbed by factors such as bone occlusion and body position changes (such as slight movement of the patient's limbs), resulting in identification fluctuations. Therefore, a medium number of retries (such as 3-5 times) is preset. Since the surgical area is less sensitive to vibration, a shorter interval time (such as 0.5-1 second) is set to improve the efficiency of the process.

[0102] Through the parameter preset logic based on the core characteristics of the clinical scene, the retry mechanism can meet the reliability requirements of high-risk operations and adapt to the efficiency requirements of regular operations, convert clinical experience into quantifiable system parameters, and realize precise matching of technical solutions and clinical needs.

[0103] If the result is "tracking mark detected", subsequent steps are continued, that is, based on the mark recognition result and the current position information, position adjustment information is obtained.

[0104] The technical scheme provided by the embodiments of the application has the following technical benefits:

[0105] 1. In the traditional scheme, a single detection failure requires manual troubleshooting by the doctor (such as repositioning the marker and adjusting the position of the tracker), which interrupts the surgical procedure. The "intelligent retry + conditional optimization" mechanism of the present application can automatically solve accidental problems such as temporary occlusion and light fluctuation, and only prompts manual intervention after multiple failures, greatly reducing human intervention during surgery and improving process continuity.

[0106] 2. In the surgical scene, the marker is temporarily occluded due to bleeding, instrument movement, and other reasons, and the traditional scheme is difficult to cope with. The retry logic of the present application combined with the dynamic optimization strategy (such as adjusting the exposure and fine-tuning the passive arm) can adapt to changes in the surgical field environment, maintain a high recognition success rate even in complex working conditions, and solve the pain point of the traditional scheme that requires high environmental stability.

[0107] 3. In the traditional scheme, the detection result and the subsequent adjustment process lack hard correlation (such as the doctor starting the adjustment without confirming the position of the marker); the rigid logic of "detection success → automatically entering the adjustment process" in the present application ensures that pose adjustment is only performed after accurate identification of the marker, avoiding positioning deviations caused by false identification, and providing double protection for surgical accuracy.

[0108] 4. Unlike simple fixed number of retries, the present application will optimize the detection parameters (exposure, matching threshold, passive arm fine-tuning angle) according to the specific reason for each undetected (such as occlusion, light, angle deviation) when retrying, forming a closed-loop logic of "failure reason → optimization measure → retry".

[0109] 5. The present application includes fine-tuning of the passive arm in the retry process (such as slightly rotating the passive arm to avoid occlusion), which enables the "passive coarse adjustment" function to play a role again in the detection stage, realizing deep collaboration between the passive shaft and the recognition algorithm, and breaking through the limitation of the traditional scheme that the passive shaft is only used for initial coarse adjustment.

[0110] 6、When multiple retries fail, the system not only alarms, but also outputs detailed logs containing the reason for each failure (such as occlusion, insufficient light, marker offset), helping doctors quickly locate the problem and providing data support for subsequent algorithm optimization. This "alarm + trace" design improves the maintainability and iteration capability of the system.

[0111] 7、According to the accuracy requirements and complexity of different surgical types (neurosurgery, orthopedics), the system presets differentiated retry times and interval times, making it suitable for various clinical scenarios and embodying the flexibility of "general architecture + scenario customization".

[0112] 103、Based on the identification recognition result and the current position information, obtain position adjustment information.

[0113] Specifically, the system first synchronizes the identification recognition result (such as the two-dimensional coordinates and attitude angle of the tracking marker in the optical tracking position instrument field of view) and the current position information (such as the three-dimensional coordinates and real-time angle of each axis of the tracking instrument in the system initial coordinate system) in time and space. Through timestamp matching, it ensures that both correspond to the state at the same time, and through coordinate reference conversion (mapping the position of the marker in the image to the system initial coordinate system), it realizes spatial alignment. For example, the identification recognition result shows that "marker is located at (320, 200) pixels in the image, tilted by 3°", and the current position information shows that "the tracking instrument is located at (1500, 300, 1800) mm in the base coordinate system, with a horizontal rotation angle of 0° and a pitch angle of 2°". After fusion, a unified description of "marker is located at (1550, 320, 1750) mm in the base coordinate system, with a total tilt of 3°" is obtained, eliminating the time and space deviation between the information sources.

[0114] Then, based on the fused information, the system constructs a dynamic deviation model of "current state-target state": the preset target state is "marker located at the center of the tracking instrument field of view and attitude level (tilt 0°)", and by calculating the difference between the current state and the target state, the deviation amount of six degrees of freedom (position deviation in X, Y, Z directions, attitude deviation around X, Y, Z axes) is quantified. For example, the fused data shows that the marker deviates from the center by 50 mm in the X direction, 20 mm in the Y direction, and 100 mm in the Z direction, and tilts by 3° around the X axis. The deviation model outputs: ΔX = +50 mm (need to move right), ΔY = +20 mm (need to move down), ΔZ = +100 mm (need to move away), Δα = +3° (need to rotate clockwise to correct the tilt), Δβ = 0°, Δγ = 0°. This model not only contains static deviation, but also analyzes the deviation change rate (such as marker moving at 2 mm / s in the positive direction of X axis) through continuous frame comparison, providing dynamic basis for adjustment amount generation.

[0115] Then, the system combines the surgical scene constraints (such as neurosurgery needs to avoid rapid movement, orthopedics can appropriately improve efficiency) and mechanical structure restrictions (maximum adjustment speed, acceleration of each axis), and converts the quantitative deviation in the deviation model into the final position adjustment information. For example, for a deviation of ΔX = +50mm, ΔZ = +100mm, Δα = +3°, considering the safety constraints of the neurosurgery scene, the generated position adjustment information is: "adjust the X direction in 3 steps (each step is 17mm, interval 0.5s), adjust the Z direction 100mm at a time (speed 50mm / s), rotate 3° around the X axis (speed 1° / s)", while including the sequence of each adjustment step (adjust the Z axis distance first, then correct the X axis position, and finally adjust the attitude), as much as possible to move smoothly and without interference.

[0116] The technical scheme provided by the embodiments of the present application has the following technical benefits:

[0117] 1. The traditional scheme only calculates the adjustment amount based on a single information source (such as only using image recognition results), which is easily affected by local errors. The present application fuses the identification recognition and the current position information through space-time alignment, eliminates the deviation caused by information islands, and makes the adjustment basis more comprehensive.

[0118] 2. The traditional scheme only calculates the static deviation and cannot cope with the marker movement during the operation (such as body movement caused by patient breathing). The present application constructs a dynamic model through deviation change rate analysis, and the generated adjustment information includes a pre-judgment compensation amount (such as adjusting 2mm in advance to offset the movement trend of the marker), which significantly improves the adaptability of complex operation areas.

[0119] 3. The traditional scheme uses fixed adjustment parameters and cannot meet the needs of different types of operations. Scheme 6 dynamically optimizes the adjustment speed, steps and sequence according to the scene constraints, and prioritizes safety (low speed, step adjustment) in sensitive scenes such as neurosurgery, and considers efficiency (fast, continuous adjustment) in scenes such as orthopedics.

[0120] Further, in some embodiments, based on the identification recognition result and the current position information, the position adjustment information is obtained, including:

[0121] Based on the visual perception model, the identification recognition result is coordinate mapped to obtain a first coordinate system transformation relationship between the optical tracking position instrument coordinate system and the tracking identification coordinate system;

[0122] The first coordinate system transformation relationship and the current position information are coordinate fused to obtain a second coordinate system transformation relationship between the system initial coordinate system and the tracking identification coordinate system;

[0123] Based on the motion control algorithm, the second coordinate system transformation relationship is calculated and parameter solved for pose deviation to obtain the position adjustment information.

[0124] Specifically, first, the identification recognition result (tracking the position and posture of the identification in the image) is converted into the spatial position relationship between the optical tracking position instrument (camera) and the tracking identification. The visual perception model integrates the binocular vision principle and the marker recognition algorithm, and calculates the three-dimensional distance and direction of the tracking identification relative to the camera by analyzing the image parallax collected by the binocular camera. For example, in a neurosurgery operation, after the tracking identification (marker pasted on the patient's skull) is captured by the camera, the system calculates the specific position of the marker in the camera coordinate system (such as 800mm in front of the camera, 50mm to the left, and 30mm below) by identifying the position difference of the corner points of the marker in the left and right images, and determines the posture of the marker (such as rotating 5° around the Z axis of the camera) according to the tilt angle of the marker. These position and posture parameters jointly constitute the first coordinate system transformation relationship between the optical tracking position instrument coordinate system and the tracking identification coordinate system, which directly reflects the relative spatial state of the two.

[0125] Then, the relative relationship between the camera and the tracking identification is incorporated into the unified system reference (system initial coordinate system, i.e. robot base coordinate system), realizing the reference unification of all position information. The current position information has clearly indicated the position and posture of the camera in the system initial coordinate system (such as 1500mm in front of the base, 300mm to the right, 1800mm in height, and 10° tilt around the X axis), combined with the above-mentioned "relative relationship between the camera and the marker", through the coordinate system theory (stacking the position parameters in different coordinate systems through the transformation matrix), the position of the marker in the system initial coordinate system can be calculated. For example, the camera is located at (1500, 300, 1800) in the base coordinate system, and the marker is located at (-50, 30, 800) in the camera coordinate system, then through coordinate fusion, the position of the marker in the base coordinate system is (1500-50, 300+30, 1800+800) = (1450, 330, 2600), and the posture parameters of the two are stacked, finally forming the second coordinate system transformation relationship between the system initial coordinate system and the tracking identification coordinate system, providing a global unified coordinate reference for subsequent adjustment.

[0126] Then, the algorithm calculates the deviation of the tracking mark between the current pose and the target pose, and converts it into specific adjustment parameters of each active shaft. The motion control algorithm first presets the target pose of the tracking mark (such as the ideal position and attitude in the initial coordinate system of the system, or the spatial position corresponding to the center of the camera field of view), and then compares the current pose obtained above with the target pose to calculate the deviation value (such as position deviation X = 50 mm, Y = -30 mm, Z = 100 mm, and attitude deviation 5° around the X axis and 3° around the Y axis). Then, the deviation values are distributed to each active shaft through parameter solving (inverse kinematics solving): for example, the Z axis deviation of 100 mm corresponds to the height adjustment data of the lifting arm (raising 100 mm), the attitude deviation of 5° around the X axis corresponds to the first angle adjustment data of the horizontal rotating arm (rotating 5°), and the attitude deviation of 3° around the Y axis corresponds to the second angle adjustment data of the pitching arm (pitching 3°). The final output position adjustment information directly guides the coordinated action of each active shaft, ensuring that the tracking mark accurately reaches the target pose.

[0127] The technical scheme provided by the embodiments of the present application has the following technical benefits:

[0128] 1. The traditional scheme relies on preset path adjustment and cannot cope with the slight movement of intraoperative markers or patients; the present application updates the mark recognition result and current position information in real time, dynamically calculates the deviation and generates adjustment parameters, so that even if the marker is slightly displaced (such as body movement caused by patient breathing) during the operation, it can quickly respond and correct, ensuring accurate positioning at all times.

[0129] 2. In the traditional scheme, each shaft adjustment is performed independently, which is prone to motion conflicts (such as out-of-sync lifting and rotating, causing the field of view to shift); the present application uniformly distributes the deviation values to each active shaft through parameter solving, ensuring that the actions of the lifting arm, the horizontal rotating arm, and the pitching arm are based on the same deviation reference, realizing coordinated linkage and avoiding loss of the field of view during adjustment.

[0130] 3. The traditional scheme is sensitive to changes in the surgical area environment (such as occlusion and changes in light), and the adjustment is prone to failure; the present application uses dynamic identification of the visual perception model and global calculation of coordinate fusion, so that even if there is slight interference in the surgical area, it can ensure that the adjustment direction is correct through accurate deviation calculation, greatly improving the stability of the system in complex scenarios.

[0131] 104. Obtain height adjustment data, first angle adjustment data, and second angle adjustment data based on the position adjustment information.

[0132] Specifically, the position adjustment information contains the comprehensive deviation (e.g., the sum of spatial position deviation and attitude deviation) between the current pose and the target pose of the optical tracking positioner. The system first decomposes this comprehensive deviation into translation deviations (ΔX, ΔY, ΔZ) along the three coordinate axes and rotation deviations (Δα, Δβ, Δγ) around the three coordinate axes. For example, if the position adjustment information shows that the tracking marker is deviated to the left and up in the field of view and is too close to the target, it is analyzed that the translation deviations are ΔX = +30 mm (need to move to the right), ΔY = +20 mm (need to move down), and ΔZ = +50 mm (need to move away from the target); the rotation deviations are Δα = +5° (clockwise rotation around the X axis to correct the horizontal tilt), Δβ = +3° (upward rotation around the Y axis to correct the vertical tilt), and Δγ = 0° (no need to rotate around the Z axis). This step converts the abstract adjustment requirement into concrete spatial parameters by establishing a mapping relationship between the deviation and the physical motion.

[0133] Then, according to the functional characteristics of each active axis (the Z-axis translation corresponds to the lifting axis, the rotation around the X axis corresponds to the horizontal rotation axis, and the rotation around the Y axis corresponds to the pitch axis), the analyzed deviations are assigned to the corresponding axis system. The system has a built-in axis system function mapping table: the lifting axis is responsible for compensating the ΔZ deviation, the horizontal rotation axis is responsible for compensating the Δα deviation, and the pitch axis is responsible for compensating the Δβ deviation; for the ΔX and ΔY translation deviations, they are indirectly realized through the coordinated rotation of the horizontal rotation axis and the pitch axis (using the horizontal / vertical projection displacement generated by the rotation motion). For example, ΔX = +30 mm is realized by rotating the horizontal rotation axis clockwise by 5° (corresponding to a horizontal direction projection displacement of 30 mm), ΔY = +20 mm is realized by rotating the pitch axis upward by 3° (corresponding to a vertical direction projection displacement of 20 mm), and ΔZ = +50 mm is directly realized by lifting the lifting axis by 50 mm. During the assignment process, the system will real-time check the load of each axis (such as motor current) to ensure that the assigned deviation value is within the bearing range of the axis system.

[0134] The assigned adjustment parameters of each axis are subjected to precision verification and dynamic optimization: through the images captured by the optical tracking positioner in real time, the theoretical marker displacement corresponding to the adjustment parameters is calculated and compared with the expected displacement, if the deviation exceeds 0.5 mm (or 0.1°), the parameters are corrected through an iterative algorithm (such as fine-tuning the horizontal rotation axis angle by 0.2°); at the same time, combined with the scene restrictions (such as the need to reduce the movement speed in neurosurgery), the parameters are optimized for safety (such as reducing the lifting axis speed from 10 mm / s to 5 mm / s). The final output height adjustment data (such as "lift the lifting axis by 50 mm at a speed of 5 mm / s"), first angle adjustment data (such as "rotate the horizontal rotation axis clockwise by 5.2° at an acceleration of 2° / s 2 "), and second angle adjustment data (such as "rotate the pitch axis upward by 3.1° at an acceleration of 1° / s 2 "), can directly drive the motor to execute.

[0135] The technical scheme provided by the embodiments of the present application has the following technical benefits:

[0136] The traditional scheme simply allocates all deviations to corresponding shafts, resulting in excessive load on a certain shaft (for example, a large displacement is compensated by a translation shaft alone); the present application reasonably allocates the load by the collaborative logic of "rotary shaft compensating for translation deviation", reduces the movement amplitude of a single shaft, and prolongs the service life of the machine.

[0137] Further, in some embodiments, the height adjustment data, the first angle adjustment data and the second angle adjustment data are obtained based on the position adjustment information, including:

[0138] Based on the kinematics inverse solution model, the target position and target attitude parameters of the optical tracking positioner in the system initial coordinate system are obtained by decomposing the target pose parameters in the position adjustment information;

[0139] Based on the active shaft motion constraint condition, the target position and target attitude parameters are mapped to the shaft system to generate a height deviation value corresponding to the lifting shaft, a first angle deviation value corresponding to the horizontal rotation shaft, and a second angle deviation value corresponding to the pitch shaft, respectively;

[0140] Based on the servo control rule, the height deviation value, the first angle deviation value and the second angle deviation value are quantitatively calibrated to obtain the height adjustment data, the first angle adjustment data and the second angle adjustment data.

[0141] Specifically, first, the abstract "target pose parameters" are converted into quantifiable spatial position and attitude data. The target pose parameters in the position adjustment information are the final state that the system expects the optical tracking positioner to reach (for example, the position corresponding to the central position of the field of view of the tracking mark), and the kinematics inverse solution model is a reverse deduction of the specific spatial parameters corresponding to this target state through a pre-established mathematical relationship of the mechanical structure. For example, if the position adjustment information requires "moving the tracking mark from the left side of the current field of view to the center", the system calculates through the inverse solution model that in the system initial coordinate system (the base coordinate system of the mechanical arm), the target position of the optical tracking positioner needs to be moved 50mm to the right in the X-axis direction and the Z-axis height remains unchanged, and the target attitude needs to be rotated 10° clockwise around the horizontal rotation shaft (X-axis) and 5° upward around the pitch shaft (Y-axis) to ensure that the field of view is horizontally aligned and vertically centered. These decomposed target position (X, Y, Z coordinates) and target attitude (rotation angle around each shaft) parameters clearly define the spatial state that the tracking instrument needs to reach, providing an accurate basis for subsequent shaft system mapping.

[0142] Then, the target pose parameters are associated with the mechanical movement range of each active axis, and the deviation amount required for adjustment of each axis is calculated. The active axis movement constraint condition refers to the physical limit (such as the maximum travel of the lifting axis of 1500-2000 mm, the rotation range of the horizontal rotation axis of ±60°, and the rotation range of the pitch axis of ±45°) and the movement accuracy (such as the minimum adjustment amount of the lifting axis of 1 mm and the minimum adjustment amount of the angle axis of 0.1°) of each axis. For example, the height of the Z axis in the target position generated in step 331 is 1800 mm, and the current height of the Z axis of the tracker is 1750 mm. Based on the constraint condition of the lifting axis (allowing movement within the range), the height deviation value is calculated as +50 mm (need to be raised by 50 mm); the target pose needs to be rotated by 10° around the X axis, and the current angle of the horizontal rotation axis is 0°, so the first angle deviation value is +10°; the target pose needs to be rotated by 5° around the Y axis, and the current angle of the pitch axis is 2°, so the second angle deviation value is +3°. If the target parameter exceeds the constraint range (such as the target Z axis height of 2100 mm exceeding the maximum travel), the system will automatically intercept the maximum allowed value (2000 mm) and record the deviation, to ensure that the movement of each axis is within a safe range.

[0143] Next, the deviation value is converted into accurate instructions that can directly drive the motor through the control logic of the servo motor. The servo control rules include the movement speed of the motor, the acceleration limit (to avoid mechanical impact caused by sudden stop and start), closed-loop feedback calibration (such as real-time correction of position error through an encoder), and the like. For example, the height deviation value of +50 mm needs to be converted into the number of rotations of the lifting axis motor (according to the screw pitch, 50 mm corresponds to 10 rotations of the motor), and the speed is set to 5 mm / s (to avoid vibration of the tracker caused by rapid lifting); the first angle deviation value of +10° needs to be converted into the number of pulses of the horizontal rotation axis motor (10° corresponds to 1000 pulses), and the actual rotation angle is detected in real time through an encoder. If the feedback deviation is 0.2° when the rotation is 9.8°, 20 pulses are supplemented for calibration; the second angle deviation value of +3° is converted into the instruction parameters of the pitch axis motor in the same way. The final output height adjustment data (such as “rotate 10 turns, speed 5 mm / s”), the first angle adjustment data (such as “1000+20 pulses, acceleration 10° / s 2 ”), and the second angle adjustment data can be directly executed by the control module to ensure that each axis accurately reaches the target position.

[0144] The technical scheme provided by the embodiments of the present application has the following technical advantages:

[0145] 1. The traditional scheme relies on manual estimation of the adjustment amount, which is easy to exceed the safe range of the mechanical structure (such as excessive lifting of the lifting axis leading to tilting); the present application strictly limits the deviation value within the physical limit through the active axis movement constraint condition, and ensures smooth movement through the servo control rule, thereby double-protecting the safety of the operation from the mechanical and control levels.

[0146] 2、Traditional scheme each axis independent adjustment, appear "lift axis lift at the same time horizontal rotation axis over rotation caused by the loss of field of view" problem; the application is based on the same target pose parameter decomposition, mapping and calibration, ensure the lift, horizontal rotation, pitch action based on the unified adjustment reference, each axis motion matching each other, solve the traditional scheme of action uncoordinated pain points.

[0147] 3、Traditional scheme for open loop control (after issuing instructions no longer correction), easy to cause mechanical wear and tear actual position and target deviation; the application through servo control of closed loop feedback (such as encoder real-time correction), can dynamically compensate for error (such as the above angle deviation 0.2° calibration), ensure that the final adjustment data and target parameters consistent, meet the high precision requirements of surgery.

[0148] 4、Traditional scheme for mechanical arm load change (such as tracking instrument slightly shaking) or external interference sensitive, adjustment failure; the application of kinematics inverse model based on mechanical structure characteristics, constraint conditions considering the load limit, servo rules can resist small amplitude interference, so that the system in different surgical scenarios (such as neurosurgery high precision requirement, orthopedic large load adjustment) can work stably.

[0149] 105、Based on the height adjustment data to adjust the height of the lifting arm, based on the first angle adjustment data to adjust the first angle of the horizontal rotation arm, based on the second angle adjustment data to adjust the second angle of the pitch arm, so that the tracking mark is located in the preset identification position of the optical tracking positioner.

[0150] Specifically, the lifting arm is composed of a lead screw sliding table structure driven by a servo motor, and the system converts the height adjustment data (such as "raise 50mm") into the rotation instruction of the motor. After the motor starts, the displacement sensor (such as grating ruler) installed on the lifting shaft feeds back the current height in real time, and compares it with the target height (current height + 50mm). If the actual height is lower than the target value, the motor continues to rotate forward; if it exceeds, it reverses to fine tune until the deviation is less than 0.5mm (satisfying the surgical precision requirement). For example, in orthopedic surgery, if the tracking mark is blocked in the lower half of the field of view due to the low patient position, the system controls the lifting arm to raise 30mm according to the height adjustment data, so that the mark appears completely in the field of view.

[0151] The horizontal rotation arm is connected with the servo motor through the harmonic reducer and can rotate around the shaft. The system converts the first angle adjustment data (such as "rotate clockwise 10°") into the angle instruction of the motor. When the motor rotates, the angle encoder on the rotating shaft detects the current angle in real time, compares it with the target angle (current angle + 10°), and corrects the rotation speed through the PID algorithm (such as slowing down when approaching the target), to ensure that the final angle deviation is less than 0.1°. For example, in neurosurgery, if the mark is deviated to the left side of the field of view, the system controls the horizontal rotation arm to rotate clockwise by 8°, so that the mark moves to the horizontal center line of the field of view.

[0152] The pitch arm and the horizontal rotation arm are connected by a rotating shaft and are driven to rotate around the shaft by a servo motor. The adjustment logic is similar to that of the horizontal rotation arm. The second angle adjustment data (such as "pitch up 5°") is converted into motor instructions, which are fed back and corrected in real time by an angle encoder to ensure that the pitch angle deviation is less than 0.1°. For example, if the logo is deviated to the lower part of the field of view, the system controls the pitch arm to rotate upward by 6°, so that the logo moves to the vertical center line of the field of view.

[0153] During the adjustment of the three axes, the binocular camera of the optical tracking positioner synchronously captures images to determine in real time whether the logo enters the preset recognition position. If the adjustment of an axis causes the logo to deviate (for example, the angle deviates due to mechanical deformation during the lifting process), the system will immediately pause the current adjustment and recalculate the compensation adjustment data of the other axes (for example, adjust the pitch arm by 0.5° to correct the deviation), until the logo is stably positioned in the preset position (for example, within ±2 pixels of the center of the field of view).

[0154] In one embodiment, with reference to Figure 3 , the overall control process of the automatic positioning system can be executed as follows: first, the system is initialized, the system is powered on, and the automatic positioning system enters the initialization state. All active axes (lifting axis, rolling axis, and pitching axis) obtain the current position and enter the enabled state.

[0155] At this time, the doctor or surgical personnel needs to paste or place a specific marker in the surgical area according to the surgical requirements as the target area for the tracking instrument to recognize and automatically adjust.

[0156] Then, manual coarse positioning is performed. The doctor or surgical personnel manually adjusts the two passive joints according to the surgical requirements to quickly point the optical tracking positioner to the general direction of the patient's surgical area. This process does not require accurate adjustment, only a general pointing to the surgical area is needed. This process generally takes less than 10 seconds.

[0157] Next, automatic search is started. After coarse positioning, the automatic positioning system enters the automatic search marker state by sending a control signal through the software, and the marker is monitored in real time. The optical tracking positioner detects whether the marker of the surgical area appears in the field of view in real time according to the marker recognition algorithm.

[0158] Then the search path is planned, the end effector pose is calculated in real time according to forward kinematics-improved DH parameters, and the camera pose can be calculated in real time through the fixed transformation from the end effector to the camera coordinate system. According to the current tracker pose, the search path is automatically planned. The control module communicates with the motor through the CANopen protocol, decomposes the planned path into the rotation information of each axis, and sends the command to the motor to start rotating. When the marker is detected in the field of view of the binocular camera, the search process is terminated, and the automatic fine positioning process is entered.

[0159] In the automatic fine positioning process, when the binocular camera recognizes the marker in the surgical area, the algorithm automatically guides the motion of the active arm so that the marker is located at a specific center position or a user-defined position in the camera field of view. The core steps include coordinate system transformation, spatial position adjustment, and final pose recalculation and adjustment. The following is a detailed description of the automatic fine positioning algorithm:

[0160] In the first stage, the relationship between the coordinate systems is obtained, and the complete base coordinate system to marker transformation chain is constructed. First, define the coordinate system transformation matrix which represents the transformation from the base coordinate system of the robot arm to the camera coordinate system, using the homogeneous matrix representation:

[0161]

[0162] where R 3*3 represents the rotation matrix of the coordinate system, and t 3*1 represents the translation matrix of the coordinate system.

[0163] The homogeneous transformation matrix The calculation formula is:

[0164]

[0165] where represents the transformation from the base coordinate system of the robot arm to the end effector, which is calculated through the forward kinematics of the robot arm.

[0166] represents the transformation from the end effector to the camera coordinate system. Since the end effector and the camera are rigidly connected, the transformation relationship between them is fixed and obtained through hand-eye calibration in advance.

[0167] The transformation relationship from the camera coordinate system to the marker coordinate system is calculated through the binocular vision principle and the marker recognition algorithm Combined with the The transformation relationship from the robot base coordinate system to the marker coordinate system can be calculated

[0168]

[0169] The second stage is visual offset processing, input: the expected offset (Δx, Δy) of the marker in the image, unit: millimeter, output: a virtual target point in 3D space, the purpose is to make the actual marker appear at the specified position in the image when the camera is aligned.

[0170] First, the position M of the marker in the robot base coordinate system is obtained:

[0171]

[0172] Then, the position matrix C and the rotation matrix R of the camera in the base coordinate system are obtained:

[0173]

[0174] The X and Y axis vectors of the robot coordinate system (in the base coordinate system) can be obtained through the rotation matrix, defined as and The calculation method is:

[0175]

[0176]

[0177] Calculate the virtual target point M in 3D space adjusted When the camera position is adjusted to align the virtual target point, the marker appears at the specified position in the field of view:

[0178]

[0179] Where Δx and Δy represent the offset values of the marker relative to the origin of the camera coordinate system in the x and y axis directions, respectively, with units of millimeters. When Δx = 0 and Δy = 0, it means moving the marker to the z axis of the camera coordinate system, at this time it is considered that the marker is located at the central position of the camera field of view. Users can adjust the values of Δx and Δy according to the actual application scenario to achieve the function of adjusting the marker to different positions in the field of view.

[0180] Through the above formula, M adjusted can be calculated, which is a virtual point located in the "offset" direction of the actual marker position. When the camera is aligned to this point, the actual marker will be offset to the specified position in the image.

[0181] The third stage: calculate the pose of the camera aligned to M adjusted ​

[0182] Input:

[0183] M adjusted Virtual target point (3D) calculated in the second stage,

[0184] C: Current position of camera (3D),

[0185] R: Current rotation matrix of camera (3x3)

[0186] Δl: User input, distance from camera to original marker point M (scalar)

[0187] M: Position of marker in base frame of robot arm

[0188] Output: C new : New position of camera, R new : New rotation matrix of camera, Next, calculate the pose of the camera after moving, first calculate the direction of Z axis, since it is considered that only by rotating the camera can the marker be adjusted to the target position, so the origin position of the camera coordinate system in the base coordinate system does not change, but the three axis directions are rotated. Since M adjusted and C two points are located on the rotated Z axis, so the direction vector of Z axis can be determined by the two points

[0189]

[0190] Since only the Z axis is not enough to uniquely determine the spatial pose, a Y axis perpendicular to Z also needs to be determined, and the new X and Y axis directions are obtained using the Gram-Schmidt orthogonalization method

[0191] Vector and

[0192]

[0193] Where can be obtained by the formula:

[0194] .

[0195] After obtaining the three mutually orthogonal axis direction vectors, the next step is to assemble them into an orthogonal rotation matrix R that meets the right-hand rule new :

[0196]

[0197] Further construct a new transformation matrix

[0198]

[0199] where position C new is moving along the new Z axis direction by Δl (user set variable, make sure the camera and the marker are adjusted to a suitable distance) unit is millimeter.

[0200]

[0201] Finally, calculate the new end effector pose

[0202]

[0203] Then use the inverse solution of the improved DH (Modified DH) parameters to make the robot move to this pose, complete the automatic fine positioning process based on vision.

[0204] It should be understood that the present application is not limited to the processes and structures already described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. An automatic positioning method of an optical tracking positioner of a surgical robot, applied to an automatic positioning system, the automatic positioning system comprising a lifting arm, a passive arm rotationally connected with the lifting arm, a horizontal rotation arm rotationally connected with the passive arm, a pitch arm rotationally connected with the horizontal rotation arm, and an optical tracking positioner fixedly connected with the pitch arm, characterized in that, The automatic positioning method comprises: When the automatic positioning system is started, the current position information of the optical tracking positioner is acquired; An identification recognition result of a tracking mark of the optical tracking positioner is acquired; Based on the identification recognition result and the current position information, position adjustment information is obtained; Based on the position adjustment information, height adjustment data, first angle adjustment data and second angle adjustment data are acquired; Based on the height adjustment data, the height of the lifting arm is adjusted, based on the first angle adjustment data, the first angle of the horizontal rotating arm is adjusted, and based on the second angle adjustment data, the second angle of the pitching arm is adjusted, so that the tracking mark is located at the preset identification position of the optical tracking positioner.

2. The auto-positioning method of claim 1, wherein, When the automatic positioning system is started, the current position information of the optical tracking positioner is acquired, which comprises: When the automatic positioning system is started, the current state of the active joint in the automatic positioning system is parameterized based on a preset kinematic model, to obtain a transformation relationship from the system initial pose to the end pose; Based on a pre-established connection relationship model, the fixed relationship between the end and the optical tracking positioner is analyzed to obtain a transformation relationship from the end to the optical tracking positioner; Based on a preset coordinate transformation theory, the transformation relationship from the system initial pose to the end pose and the transformation relationship from the end to the optical tracking positioner are synthesized to acquire the current position information of the optical tracking positioner in the system initial coordinate system.

3. The auto-positioning method of claim 1, wherein, Based on the identification recognition result and the current position information, the position adjustment information is obtained, which comprises: Based on a visual perception model, the identification recognition result is coordinate-mapped to obtain a first coordinate system transformation relationship between the optical tracking positioner coordinate system and the tracking mark coordinate system; The first coordinate system transformation relationship and the current position information are coordinate-fused to obtain a second coordinate system transformation relationship between the system initial coordinate system and the tracking mark coordinate system; Based on a motion control algorithm, the second coordinate system transformation relationship is pose deviation calculated and parameter solved to obtain the position adjustment information.

4. The auto-positioning method of claim 1, wherein, Based on the position adjustment information, the height adjustment data, the first angle adjustment data and the second angle adjustment data are acquired, which comprises: Based on a kinematic inverse solution model, the target pose parameters in the position adjustment information are decomposed to obtain target position and target attitude parameters of the optical tracking positioner in the system initial coordinate system; Based on the active shaft motion constraint condition, the target position and target attitude parameters are shaft system mapped to generate a height deviation value corresponding to the lifting shaft, a first angle deviation value corresponding to the horizontal rotating shaft, and a second angle deviation value corresponding to the pitching shaft, respectively; Based on a servo control rule, the height deviation value, the first angle deviation value and the second angle deviation value are quantitatively calibrated to obtain the height adjustment data, the first angle adjustment data and the second angle adjustment data.

5. The auto-positioning method of claim 1, wherein, After the identification recognition result of the tracking mark of the optical tracking positioner is acquired, it further comprises: If the identification result is that the tracking mark is not detected, the detection operation on the tracking mark is re-executed for a preset number of times and then stopped; If the identification result is that the tracking mark is detected, the position adjustment information is obtained based on the identification result and the current position information.

6. The auto-positioning method of claim 1, wherein, When the automatic positioning system is started, the method further comprises: In response to an automatic search instruction, the identification recognition data of the mark in the field of view of the optical tracking positioner is obtained in real time based on a visual detection module; Based on the current position information of the optical tracking positioner in the system initial coordinate system, the current pose parameters of the optical tracking positioner are calculated in combination with a forward kinematics model; Based on the current pose parameters and a preset search range constraint condition, a search path covering the possible area of the operation area is planned, and the search path includes the movement trajectory of the optical tracking positioner in the lifting shaft direction, the rotation trajectory of the horizontal rotation shaft, and the rotation trajectory of the pitch shaft; According to the search path, the control module sends a motion instruction to each driving shaft to drive the optical tracking positioner to perform search motion, and simultaneously continuously detects whether the tracking mark appears in the field of view to obtain the identification result.

7. The auto-positioning method of claim 1, wherein, Before the current position information of the optical tracking positioner is obtained, the method further comprises: Obtaining the passive adjustment result of the passive arm; Based on the passive adjustment result, the current position information of the optical tracking positioner is obtained.

8. An automatic positioning system for applying an automatic positioning method of a surgical robot using an optical tracking localizer according to any one of claims 1 to 7, characterized in that, Comprise: a base; a lifting arm, one end of which is fixedly arranged on the base, for adjusting the height of the optical tracking positioner; a passive arm, one end of which is rotatably connected to the end of the lifting arm away from the base, and the rotation axis of the passive arm is parallel to the lifting arm; a horizontal rotation arm, one end of which is rotatably connected to the end of the passive arm away from the lifting arm, and the rotation axis of the horizontal rotation arm is perpendicular to the lifting arm; a pitch arm, one end of which is rotatably connected to the end of the horizontal rotation arm away from the passive arm, and the other end of which is fixedly connected to the optical tracking positioner, and the rotation axis of the pitch arm is perpendicular to the horizontal rotation arm.

9. The automatic positioning system of claim 8, wherein, The passive arm comprises a first arm and a second arm; one end of the first arm is rotatably connected to the lifting arm, and the other end is rotatably connected to the second arm, and one end of the second arm away from the first arm is rotatably connected to the horizontal rotation arm, and the rotation axis of the first arm is parallel to the rotation axis of the second arm.

Images