A surgical navigation system based on optical motion capture and ring LED markers

Abstract

The application relates to the technical field of surgical navigation, and particularly discloses a surgical navigation system based on optical motion capture and annular LED markers, which comprises an annular LED marker module, the annular LED marker module comprises an LED annular lamp which is composed of a plurality of infrared LEDs uniformly distributed on an annular substrate and is used for forming an infrared spot array which can be recognized by a binocular infrared camera, the annular LED provides a distributed uniform and easily recognized spot array through the design of the annular LED, and the robustness of the system to occlusion is enhanced; a weighted optimization algorithm improves the accuracy and anti-interference ability of pose solution; an integrated data analysis module provides real-time guarantee for the reliability of navigation; and a functionally perfect surgical navigation module converts abstract pose data into an intuitive anatomical relationship view, so that, under the efficient cooperation of the modules, a stable, accurate and intelligent navigation scheme for minimally invasive surgery based on optical motion capture technology can be realized, the surgical risk is effectively reduced, and the success rate of surgery is improved.

Description

Technical Field

[0001] This invention relates to the field of surgical navigation technology, specifically a surgical navigation system based on optical motion capture and ring LED markers. Background Technology

[0002] Surgical navigation systems are an important technological support for modern precision medicine. They assist doctors in performing high-precision surgical operations under the guidance of preoperative images by tracking the spatial position and posture of surgical instruments in the patient's body in real time.

[0003] Mainstream optical surgical navigation systems primarily employ active infrared LED marking schemes. In these schemes, several infrared LEDs are fixed to the surgical instrument to form a rigid marking body. An infrared camera synchronously acquires images of the LED spot. The image processing module extracts the center coordinates of the spot and reconstructs the three-dimensional spatial coordinates of each LED through stereo matching and triangulation. Based on the known geometric layout of the LEDs on the rigid body, the six-degree-of-freedom pose of the instrument is solved using least-squares fitting or iterative nearest-point algorithm. Finally, the pose is registered to the preoperative image coordinate system to achieve navigation display.

[0004] In existing technologies, traditional active infrared LED marking schemes typically assign equal weights to each LED, lacking differentiated confidence assessments based on factors such as viewing angle and emission direction. This results in a significant decrease in positioning accuracy during large-angle operations. Furthermore, tracking quality judgment largely relies on the overall reprojection error threshold, lacking statistical diagnostic capabilities for individual marker errors. This makes it impossible to effectively distinguish between system calibration drift and local anomalies, affecting troubleshooting efficiency and system reliability. Summary of the Invention

[0005] The purpose of this invention is to provide a surgical navigation system based on optical motion capture and ring LED marking, solving the following technical problems: How to maintain high accuracy in pose calculation and distinguish the specific reasons for accuracy degradation in real time under conditions of large-angle operation of surgical instruments and partial LED obstruction.

[0006] The objective of this invention can be achieved through the following technical solutions: A surgical navigation system based on optical motion capture and ring LED marking, the system comprising: The ring-shaped LED marking module, fixed to the surgical instrument, includes an LED ring light, which is composed of multiple infrared LEDs evenly distributed on a ring substrate, used to form an array of infrared light spots that can be recognized by a binocular infrared camera; The infrared camera module includes two binocular infrared cameras mounted at a fixed baseline distance for synchronously acquiring images of an infrared spot array; The image processing module is used to receive images of the infrared light spot array, extract the center sub-pixel coordinates of each LED light spot, calculate the visibility score of each LED based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera, assign confidence weights to each LED according to the visibility scores, and then solve the three-dimensional spatial position and orientation of the surgical instrument and the weighted projection error index by minimizing the weighted projection error function. The data analysis module compares the weighted projection error index with a preset error threshold range to classify the tracking status. When the tracking is judged to be abnormal, it further performs statistical analysis on the individual weighted projection error of each visible LED to distinguish between system calibration drift and local anomalies. The surgical navigation module includes a preoperative image registration unit, a spatial registration unit, an instrument pose mapping unit, a path planning and early warning unit, and a visualization display unit; The surgical navigation module is used to establish a three-dimensional spatial reference image coordinate system based on preoperative image data, spatially register the patient's actual position with the preoperative image, map the surgical instrument pose output by the image processing module to the image coordinate system in real time, and display the positional relationship of the surgical instruments relative to the anatomical structure in a visual manner in real time.

[0007] Furthermore, the processing procedure of the image processing module includes: The three-dimensional spatial position and orientation of the surgical instrument, as well as the weighted projection error exponent, are solved by minimizing the weighted projection error function, as shown in the following formula: ; in, To add a weighted projection error index, 3 3. Rotation matrix, representing the orientation of the surgical instruments relative to the world coordinate system. 3 1. Translation vector, representing the spatial position of the surgical instrument in the world coordinate system; S, the set of visible LED points, which is the set of all LEDs successfully identified by the binocular infrared camera system at the current moment; i, an integer index, representing the number of the visible LED. Let be the confidence weight for the i-th LED, with a value between 0 and 1. Let i be the known three-dimensional coordinates of the i-th LED in the world coordinate system. Let i be the observed three-dimensional coordinates of the i-th LED in the camera coordinate system. Let be the projection transformation function. This is the symbol for the Euclidean norm.

[0008] Furthermore, the analysis process of the data analysis module includes: By weighted projection error index Compared with the preset error threshold range Perform a comparison; like If the system determines that the tracking is lost, it will automatically freeze the navigation screen, stop outputting pose data to the surgical navigation module, and issue an audible and visual alarm signal. like If the tracking is deemed abnormal, an alert will be issued; like It was judged to be a high-quality track.

[0009] Furthermore, the analysis process of the data analysis module also includes: When an anomaly is identified in the tracking, a more detailed analysis is performed based on the weighted projection error index: S1: Calculate the individual weight projection error index of each visible LED at preset time intervals during the operation; S2: Construct an error distribution vector based on the individual weight projection error index of each visible LED, and calculate the mean, standard deviation, and maximum deviation of the error distribution vector; S3: By comparing the mean, standard deviation, and maximum deviation of all visible LEDs with preset mean threshold, standard deviation threshold, and maximum deviation threshold respectively, and if any parameter is higher than the threshold, it is determined that the system calibration accuracy has exceeded the tolerance, the pose output is automatically paused and a recalibration prompt is sent to the operator.

[0010] Furthermore, the image processing procedure of the image processing module also includes: First, the visibility score of the i-th LED is calculated based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera and a preset angular constant. Then, the confidence weight of the i-th LED is calculated based on the visibility score.

[0011] Furthermore, the two binocular infrared cameras in the infrared camera module satisfy the following accuracy constraints: The depth positioning error is equal to the square of the target distance divided by the product of the baseline distance and the equivalent focal length, and then multiplied by the parallax matching accuracy. The baseline distance must be no less than the square of the maximum working distance divided by the equivalent focal length, and then multiplied by the ratio of the parallax matching accuracy to the maximum permissible depth positioning error.

[0012] Furthermore, the ring-shaped LED marker module also includes a constant current drive circuit and a microcontroller; The ring-shaped LED marking module includes 12 infrared LEDs, which are connected in a 3-series-4-parallel configuration, meaning that every 3 LEDs are connected in series as a group, and a total of 4 groups are connected in parallel, controlled by an independent constant current drive channel. The microcontroller receives commands from the host computer via the UART interface at a baud rate of 115200bps, supports 256 levels of PWM brightness adjustment and three working modes: constant light, low-frequency flicker, and high-frequency modulation, and the PWM dimming frequency is 10kHz.

[0013] Furthermore, the surgical navigation module includes: The preoperative image registration unit is used to read and reconstruct 3D CT image data and define the image coordinate system. The spatial registration unit achieves spatial mapping between the patient's actual position and the image coordinate system through point cloud matching or marker point registration algorithms. The instrument pose mapping unit is used to convert the surgical instrument pose output by the image processing module to the image coordinate system in real time. The path planning and early warning unit is used to display the preset surgical path and issue an audible and visual alarm signal when the surgical instrument navigation and tracking is lost; The visualization display unit presents the spatial positional relationship between surgical instruments and the patient's anatomical structure in real time on the display terminal in a multi-view manner.

[0014] The beneficial effects of this invention are: (1) The present invention provides a uniformly distributed and easily identifiable light spot array through the design of ring LEDs, which enhances the robustness of the system against occlusion; the weighted optimization algorithm improves the accuracy and anti-interference ability of pose calculation; the integrated data analysis module provides real-time assurance for the reliability of navigation; and the fully functional surgical navigation module transforms the abstract pose data into an intuitive view of anatomical relationships. With the efficient collaboration of each module, a stable, accurate and intelligent navigation scheme based on optical motion capture technology can be realized for minimally invasive surgery, which effectively reduces surgical risks and improves the success rate of surgery.

[0015] (2) The present invention uses a weighted optimization algorithm to minimize the reprojection error to solve the device pose. Compared with the direct use of all light spots for least square fitting or simple averaging, it has significant advantages. The introduction of confidence weights enables the algorithm to adaptively adjust according to the actual observation quality of each LED, effectively suppressing the negative impact of abnormal observations caused by ambient light interference, device aging or local occlusion on the final pose solution. This weighted optimization strategy significantly improves the positioning accuracy, anti-interference ability and overall robustness of the system in complex operating room environments, ensures the smoothness and reliability of the output pose data, and ensures the accuracy of the entire navigation system.

[0016] (3) By setting a dual threshold error range, this invention establishes a simple and efficient real-time navigation tracking quality grading evaluation system. This mechanism can transform abstract mathematical errors into intuitive three-level status indicators of "high quality", "abnormal" and "lost", which makes it easy for operators to quickly grasp the working status of the system. For the most serious "tracking loss" situation, the system is designed with multiple safety interlocking measures such as automatic screen freezing, data output suspension and sound and light alarms, which eliminates the risk of navigation screen errors caused by tracking failure and thus misleading doctors from the source, and greatly improves the safety and reliability of the surgical navigation system.

[0017] (4) By introducing statistical analysis of individual spot errors, this invention achieves an intelligent leap from “knowing the tracking anomaly” to “analyzing why it is abnormal”. This method can effectively distinguish between overall system calibration drift and local abnormal interference. The diagnostic information it provides in three dimensions—mean, standard deviation, and maximum deviation—provides a scientific basis for operators or maintenance personnel to quickly locate the root cause of the problem. Its function of automatically pausing output and prompting recalibration actively avoids navigation risks caused by accuracy degradation at the system level, further enhancing the reliability, maintainability, and clinical applicability of the system in long-term use or complex environments.

[0018] (5) This invention constructs a closed-loop accuracy assurance system of "accurate calculation → real-time evaluation → intelligent diagnosis" by organically combining a confidence-weighted algorithm based on emission angle with a multi-dimensional error diagnosis mechanism based on statistics. The weighted optimization algorithm improves the accuracy of pose calculation from the source, while the data analysis module uses the individual error data generated during the weighted optimization process for in-depth diagnosis. The two form a relationship of data reuse and functional complementarity, realizing the dual technical effects of "improved calculation accuracy" and "distinguishing the root causes of anomalies". Its overall technical effect is significantly better than the sum of the individual effects of each technical feature. Attached Figure Description

[0019] The invention will now be further described with reference to the accompanying drawings.

[0020] Figure 1 This is a schematic block diagram of a surgical navigation system based on optical motion capture and ring LED marking in this invention; Figure 2 This is a structural block diagram of the surgical navigation module in this invention; Figure 3 This is a flowchart of the data analysis module of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 and Figure 2 As shown, in one embodiment, this application provides a surgical navigation system based on optical motion capture and ring LED markers, the system comprising: The ring-shaped LED marking module, fixed to the surgical instrument, includes multiple infrared LEDs evenly distributed on the ring substrate, used to form an array of infrared light spots that can be recognized by a binocular infrared camera; The infrared camera module includes two binocular infrared cameras mounted at a fixed baseline distance for synchronously acquiring images of an infrared spot array; The image processing module is used to receive images of the infrared light spot array, extract the center sub-pixel coordinates of each LED light spot, calculate the visibility score of each LED based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera, assign confidence weights to each LED according to the visibility scores, and then solve the three-dimensional spatial position and orientation of the surgical instrument and the weighted projection error index by minimizing the weighted projection error function. The data analysis module compares the weighted projection error index with a preset error threshold range to classify the tracking status. When a tracking anomaly is detected, the module further performs statistical analysis on the individual weighted projection error of each visible LED to distinguish between system calibration drift and local anomalies. The data analysis module generates corresponding status codes based on the classification results. The status codes include at least a high-quality tracking code, an anomaly warning code, and a tracking loss code. The status codes, along with the current pose data, are sent to the surgical navigation module. The surgical navigation module includes a preoperative image registration unit, a spatial registration unit, an instrument pose mapping unit, a path planning and early warning unit, and a visualization display unit; The surgical navigation module is used to establish a three-dimensional spatial reference image coordinate system based on preoperative image data, spatially register the patient's actual position with the preoperative image, map the surgical instrument pose output by the image processing module to the image coordinate system in real time, and display the positional relationship of the surgical instruments relative to the anatomical structure in a visual manner in real time. In addition, it is also used to perform corresponding navigation response operations according to the received status code: when a tracking loss code is received, the navigation screen is frozen and the pose data update is paused; when an abnormal warning code is received, while maintaining the pose update, an accuracy abnormality prompt is superimposed on the display interface. Through the above technical solution, this embodiment provides a surgical navigation system based on optical motion capture and ring LED marking. Its working principle is as follows: after the ring LED marking module, fixed to the surgical instrument, is powered on, multiple infrared LEDs uniformly distributed on its substrate simultaneously emit infrared light of specific wavelengths, forming a spatially distributed array of infrared light spots. Based on this, two binocular infrared cameras in the infrared camera module synchronously acquire images of this light spot array with a fixed baseline, obtaining two two-dimensional images from different viewpoints at the same time. Subsequently, the image processing module preprocesses these two images, extracting each light spot. The sub-pixel coordinates of the center are used to match the corresponding light spots in the left and right images using the principle of binocular stereo vision. The coordinates of each light spot in three-dimensional space are calculated by weighted optimization algorithm, and then the six degrees of freedom (position and attitude) pose data of the entire instrument are solved. The data analysis module monitors the accuracy and reliability of the pose calculation process in real time. Finally, the surgical navigation module performs spatial registration and fusion of the pose data with the preoperative planned image coordinate system to guide the doctor's operation in a visual way. The whole system forms a complete closed-loop workflow from "light spot marking → image acquisition → pose calculation → accuracy monitoring → navigation display". This configuration, utilizing a ring-shaped LED design to provide a uniformly distributed and easily identifiable array of light spots, enhances the system's robustness against occlusion. A weighted optimization algorithm improves the accuracy and anti-interference capability of pose calculation. An integrated data analysis module provides real-time assurance for navigation reliability. A fully functional surgical navigation module transforms abstract pose data into an intuitive view of anatomical relationships. Through the efficient collaboration of these modules, a stable, precise, and intelligent navigation solution based on optical motion capture technology can be achieved for minimally invasive surgery, effectively reducing surgical risks and increasing the success rate. It should be noted that the LED ring light housing in the ring LED marking module is made of medical-grade aluminum alloy material with anodized surface treatment and a sealing rating of IP67. It can withstand high temperature and high pressure steam sterilization at 134℃. Its back is equipped with an M6 standard threaded mounting interface and a spring contact electrical connector. The spring contacts achieve electrical conduction while mechanically installing. It should be noted that the binocular infrared camera uses a global shutter CMOS image sensor with an effective pixel count of 2048×1536, a quantum efficiency of no less than 60% at a wavelength of 850nm, a frame rate of no less than 60fps, a baseline distance of 300mm, a working distance of 500-2000mm, a positioning accuracy of 0.1-0.3mm, an optical focal length of 8mm, and a physical size of 5.86μm for a single pixel of the image sensor. Each camera module is equipped with a near-infrared enhanced coated lens with a lens distortion coefficient of less than 0.5% and a narrow bandpass filter with a center wavelength of 850nm and a full width at half maximum (FWHM) of less than 50nm.

[0023] The processing procedure of the image processing module includes: The three-dimensional spatial position and orientation of the surgical instrument, as well as the weighted projection error exponent, are solved by minimizing the weighted projection error function, as shown in the following formula: ; in, The weighted projection error exponent is a scalar value representing the overall fitting residual between the current solved pose and the observed data. The smaller the value, the more accurate the solved pose. 3 3. A rotation matrix representing the orientation of the surgical instruments relative to the world coordinate system. The matrix elements are dimensionless and satisfy orthogonal constraints. 3 1. Translation vector, representing the spatial position of the surgical instrument in the world coordinate system; S, the set of visible LED points, which is the set of all LEDs successfully identified by the binocular infrared camera system at the current moment, denoted as M, where M is greater than 3; i, an integer index representing the number of the visible LED, ranging from 1 to N; and N, the total number of LEDs in the ring LED marking device. Let be the confidence weight of the i-th LED, with a value between 0 and 1, and let the weights of all visible LEDs satisfy the normalization condition. The known three-dimensional coordinates of the i-th LED in the world coordinate system are obtained by precise calibration of the mechanical design dimensions of the ring-shaped LED marker. To represent the observed three-dimensional coordinates of the i-th LED in the camera coordinate system, it is calculated using sub-pixel matching of the left and right camera images and the principle of triangulation. The projection transformation function represents the transformation of points in the camera coordinate system based on the current R and T coordinates. Mapping to the world coordinate system The symbol for the Euclidean norm represents the L2 magnitude of the vector within the brackets, which in this case represents the linear distance between two points in space. Through the above technical solution, this embodiment provides the processing procedure of the image processing module. Its working principle is based on the classic optimization idea of ​​minimizing reprojection error. Specifically, the system knows the precise three-dimensional coordinates of each LED on the ring LED mark in the world coordinate system (i.e., the local coordinate system of the surgical instrument). Simultaneously, through binocular camera observation and triangulation, the observed three-dimensional coordinates of these LEDs in the camera coordinate system at the same moment can be obtained. The image processing module needs to solve for an optimal rotation matrix R and translation vector T (describing the instrument's position) such that all observation points... go through After transforming and projecting to the world coordinate system, it is compared with the known world coordinate system. The process of minimizing the sum of squared weighted Euclidean distances between the given elements is a typical nonlinear least squares problem, usually solved using iterative optimization algorithms such as the Levenberg-Marquardt algorithm. The confidence weights introduced in the formula... This allows LEDs with higher observation quality (such as more stable emission and better viewing angle) to contribute more to the final result, thereby improving the robustness and solution accuracy of the algorithm when some LEDs are blocked or affected by noise. By employing this setup, the instrument pose is solved using a weighted optimization algorithm that minimizes reprojection error. This approach offers significant advantages over simply using all light spots for least-squares fitting or simple averaging. Furthermore, the introduction of confidence weights allows the algorithm to adaptively adjust based on the actual observation quality of each LED, effectively suppressing the negative impact of abnormal observations caused by ambient light interference, device aging, or local occlusion on the final pose calculation results. This weighted optimization strategy significantly improves the system's positioning accuracy, anti-interference capability, and overall robustness in complex operating room environments, ensuring the smoothness and reliability of the output pose data and guaranteeing the overall accuracy of the navigation system.

[0024] The analysis process of the data analysis module includes: By weighted projection error index Compared with the preset error threshold range Perform a comparison; like If the system determines that the tracking is lost, it will automatically freeze the navigation screen, stop outputting pose data to the surgical navigation module, and issue an audible and visual alarm signal. like If the tracking is deemed abnormal, an alert will be issued; like This is judged as high-quality tracking; Through the above technical solution, this embodiment uses the weighted projection error index. Compared with the preset error threshold range The core principle of the comparison is the error threshold range. This represents the system's tolerance range for positioning accuracy under different application scenarios. This is achieved through real-time calculation... By dynamically comparing the current tracking status with the specified interval, the system can quickly and automatically classify it into three levels: "high-quality tracking," "abnormal tracking," and "tracking loss." When "tracking loss" is detected, the system will immediately take safety measures to prevent the transmission of incorrect pose information to the doctor. These measures include freezing the navigation screen, stopping data output, and issuing an alarm to ensure surgical safety. This real-time monitoring and graded early warning mechanism is an important manifestation of the system's intelligence and safety. By setting up dual threshold error ranges, a simple and efficient real-time navigation tracking quality grading evaluation system was established. This mechanism can transform abstract mathematical errors into intuitive three-level status indicators of "high quality", "abnormal", and "lost", making it easy for operators to quickly grasp the system's working status. For the most serious "tracking loss" situation, the system is designed with multiple safety interlocking measures such as automatic screen freezing, data output suspension, and audible and visual alarms, which eliminates the risk of navigation screen errors caused by tracking failure and thus misleading doctors from the source, greatly improving the safety and reliability of the surgical navigation system. Based on this, a corresponding status code is generated according to the classification judgment result. The status code includes at least a high-quality tracking code, an abnormal warning code, and a tracking loss code. The status code, along with the current pose data, is sent to the surgical navigation module. The status code is a 2-bit binary integer, where 00 represents high-quality tracking, 01 represents an abnormal warning, and 10 represents tracking loss. After the data analysis module completes the classification judgment, it encapsulates the status code and the current pose data into the same data frame and sends it to the surgical navigation module via the TCP / IP protocol. After receiving the data frame, the surgical navigation module first parses the status code field and performs corresponding operations according to its value: if the status code is 00, the navigation screen is updated normally; if it is 01, the navigation screen is updated normally while displaying a yellow triangle warning icon in the upper left corner of the interface; if it is 10, the navigation screen is immediately frozen and an audible and visual alarm is triggered. This real-time, automated quality monitoring and safety response mechanism is a key guarantee for the clinical application of intelligent surgical navigation systems. It should be noted that the preset error threshold range The calibration method is as follows: First, after the system completes its factory calibration, 500 frames of static pose data are continuously acquired under standard working conditions (working distance 1000mm, ambient illuminance <100 lux, no strong infrared interference sources), and the weighted projection error of each frame is calculated. ; Then, statistical analysis was performed on the 500 frames of data collected to calculate the weighted projection error. mean with standard deviation ,Pick , In this embodiment, based on actual measurements Square pixel, Square pixels, therefore Square pixel, Square pixel.

[0025] Please see Figure 3 As shown, the analysis process of the data analysis module further includes: When an anomaly is identified in the tracking, a more detailed analysis is performed based on the weighted projection error index: S1: During the operation, at preset time intervals, calculate the individual weight projection error index of each visible LED, as shown in the following formula: ; in, Let be the weight projection error index of the i-th visible LED; S2: Construct an error distribution vector based on the individual weight projection error index of each visible LED. And calculate the mean, standard deviation, and maximum deviation of the error distribution vector; The formula for calculating the mean of the error distribution vector is as follows: ; The formula for calculating the standard deviation of the error distribution vector is as follows: ; The formula for calculating the maximum deviation of the error distribution vector is as follows: ; in, This represents the average level of the projection error of all visible LEDs, reflecting the overall accuracy of the system calibration. This indicates the degree of dispersion of the individual LED error relative to the mean, reflecting the consistency of the system calibration error at different spatial locations. This represents the maximum value of the projection error of each individual LED among all visible LEDs, reflecting the worst local condition of the system calibration accuracy; S3: By comparing the mean, standard deviation, and maximum deviation of all visible LEDs with the preset mean threshold, standard deviation threshold, and maximum deviation threshold respectively, and if any parameter is higher than the threshold, it is determined that the system calibration accuracy has exceeded the tolerance, the pose output is automatically paused and a recalibration prompt is sent to the operator. Specifically, the mean threshold is preferably 0.5 square pixels, the standard deviation threshold is preferably 0.3 square pixels, and the maximum deviation threshold is preferably 1 square pixel. When any parameter exceeds the threshold, it is determined that the system calibration accuracy has exceeded the tolerance, the current pose data output is immediately frozen, and the sending of new pose information to the navigation software system is stopped. The navigation software system keeps displaying the last frame of valid pose and prompts the operator that the current tracking has been paused in a semi-transparent gray or flashing manner. Through the above technical solution, this embodiment provides a detailed analysis process based on the weighted projection error index when a tracking anomaly is determined. Its working principle is that when the overall system error is in the abnormal range, it indicates that the positioning accuracy has decreased but has not completely failed. At this time, the root cause of the positioning accuracy decrease needs to be identified. Step S1 calculates the individual weighted projection error of each visible LED. The overall error is decomposed into each observation point. Step S2 then performs statistical analysis on these individual errors, calculating their mean, standard deviation, and maximum deviation, which are used to characterize the overall level, dispersion, and worst-case scenario of the error, respectively. Step S3 compares these three statistics with their respective preset thresholds. The core of this diagnostic logic lies in distinguishing between two sources of error: if all individual errors are large but evenly distributed (high mean, low standard deviation), it may be due to a drift in the overall calibration parameters of the system; if only a few LEDs have extremely large errors (high maximum deviation), it may be due to damage to that LED or partial obstruction. Through this fine-grained analysis, the system can provide more targeted prompts (such as "recalibration is recommended" rather than a simple alarm), providing clear guidance for troubleshooting and system maintenance. By setting it up in this way and introducing statistical analysis of individual spot errors, an intelligent leap from "knowing the tracking anomaly" to "analyzing why it is abnormal" is achieved. This method can effectively distinguish between overall system calibration drift and local abnormal interference. The diagnostic information it provides in three dimensions—mean, standard deviation, and maximum deviation—provides a scientific basis for operators or maintenance personnel to quickly locate the root cause of the problem. Its function of automatically pausing output and prompting recalibration actively avoids navigation risks caused by accuracy degradation at the system level, further enhancing the reliability, maintainability, and clinical applicability of the system in long-term use or complex environments. It should be noted that the calibration methods for the preset mean threshold, standard deviation threshold, and maximum deviation threshold are as follows: First, after the system completes its factory calibration, under standard working conditions (working distance 1000mm, ambient illuminance <100 lux, no strong infrared interference sources), 500 frames of static pose data are continuously collected, and the individual weight projection error of each visible LED in each frame is calculated. ; Then, statistical analysis was performed on all 500 frames of data, and the following calculations were performed: Average value of each frame overall mean and standard deviation Take the mean threshold as ; Standard deviation of each frame overall mean and standard deviation The standard deviation threshold is taken as ; Maximum value of each frame overall mean and standard deviation Take the maximum deviation threshold as ; In this embodiment, the measured mean threshold is 0.5 square pixels, the standard deviation threshold is 0.3 square pixels, and the maximum deviation threshold is 1.0 square pixels.

[0026] The image processing procedure of the image processing module also includes: First, the visibility score of the i-th LED is calculated based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera, and a preset angular constant, as shown in the following formula: ; Then, the confidence weight of the i-th LED is calculated based on the visibility score, as shown in the following formula: ; Where j is an integer index, representing the number of any LED in the set of visible LED points S. Let be the normalized denominator, representing the sum of the visibility scores of all LEDs in the visible LED point set S. It is a natural exponential function. Let be the angle between the normal direction of the emitting surface of the i-th LED and the optical axis of the camera, which is determined by the mounting orientation of the LED on the annular substrate and its spatial position relative to the camera. The preset angle constant; Through the above technical solution, this embodiment provides the confidence weight of the i-th LED. The calculation process works on the angle between the normal direction of the LED's emitting surface and the optical axis of the camera. The larger the LED, the weaker the effective light intensity received by the camera (following Lambert's cosine law), resulting in reduced brightness and shape distortion of the LED spot in the image, which in turn affects the sub-pixel accuracy of its center positioning. Based on this, this solution first calculates the angle between the normal of the emitting surface of each LED and the camera's optical axis, according to the installation orientation and current spatial position of each LED. Then, through a Gaussian function The visibility score of the LED is calculated, and this scoring function has a clear physical intuition: when The highest score is 1 when the angle is 0° (facing the camera directly); as... As the angle increases, the score decays exponentially, simulating the weakening of illumination as the angle increases. Finally, the score is normalized across all visible LEDs to obtain the confidence weight for each LED. This achieves dynamic and objective allocation of weights; By setting it up in this way, the contribution of each LED to the pose calculation is adjusted in real time according to the actual visible quality of each LED at the current moment (i.e. the potential light intensity to be effectively received by the camera). This allows the pose calculation process to adaptively "rely" on those LEDs that are at the best viewing angle and have the best signal quality, while "weakening" the influence of those LEDs that have too large a viewing angle and weak signal. This adaptive weighting strategy based on the physical model can significantly improve the positioning accuracy and robustness of the system when the surgical instrument is rotated at a large angle or when some LEDs are blocked. It is an ingenious and crucial part of the algorithm design. It should be noted that the preset angle constant The value of can be determined through the following simple experiment: A single infrared LED is fixed on an angle measuring stage, and an infrared camera is fixed 500mm away from the LED, with the LED normal aligned with the camera's optical axis. The center gray value of the LED spot image is recorded at this point. Then, gradually rotate the angle measuring stage to increase the angle between the LED normal and the optical axis from 0° to 70°, and record the center gray value at each angle. ,when When the value is less than 0.607 for the first time, the corresponding angle is the measured value of τ.

[0027] The two binocular infrared cameras in the infrared camera module satisfy the following accuracy constraints: The depth positioning error is equal to the square of the target distance divided by the product of the baseline distance and the equivalent focal length, and then multiplied by the parallax matching accuracy, as shown in the following formula: ; in, The depth positioning error, expressed in millimeters, represents the theoretical positioning uncertainty of the system in the depth direction. The smaller the value, the higher the depth positioning accuracy of the system. The target depth, in millimeters, represents the distance from the tracked LED to the midpoint of the baseline of the binocular infrared camera system. In this system, the value of Z ranges from 500mm to 2000mm. B is the baseline distance, in millimeters, representing the straight-line distance between the optical centers of the two cameras in a binocular infrared camera system. In this system, B = 300 mm. The equivalent focal length of a camera, expressed in pixels, is defined as follows: ,in Where is the optical focal length of the lens, which is 8mm in this system; p is the physical size of a single pixel on the image sensor, which is 5.86μm in this system. The calculated values ​​are... Pixels The disparity matching accuracy, expressed in pixels, represents the estimation accuracy of the stereo matching algorithm for the disparity value (i.e., the difference in the x-coordinate of the image position of the same spatial point in the left and right images). This system achieves this through a sub-pixel interpolation algorithm. Pixel; The baseline distance must be no less than the square of the maximum working distance divided by the equivalent focal length, and then multiplied by the ratio of the parallax matching accuracy to the maximum permissible depth positioning error, as shown in the following formula: ; in, For maximum working distance, This represents the maximum permissible depth positioning error. Through the above technical solution, this embodiment provides the accuracy constraint relationship satisfied by the two binocular infrared cameras in the infrared camera module. Its working principle is based on the classical geometric model of binocular stereo vision, the first formula. Reveals depth-direction positioning error Quantitative relationship with several key hardware parameters: error versus target distance It is proportional to the square of the distance from the baseline. and equivalent focal length The product is inversely proportional to the disparity matching accuracy. The direct proportionality, this quantitative relationship provides clear guidance for system designers, the second formula This is within a given maximum working distance. Equivalent focal length Parallax matching accuracy and maximum permissible depth error Under the premise of ensuring depth accuracy, the minimum baseline distance that must be met is calculated. By selecting parameters such as B=300mm, this system ensures that the depth positioning accuracy can theoretically reach 0.1-0.3mm within a working distance of 500-2000mm. By adopting this configuration and introducing a quantitative accuracy model based on binocular stereo vision, a scientific theoretical basis and clear constraints are provided for the design of the core hardware parameters (baseline distance, lens focal length, and pixel size) of the infrared camera module. This physical model-based design methodology ensures that the system has the theoretical foundation to achieve the designed positioning accuracy from the outset, rather than blindly selecting components or relying on subsequent software compensation. The formulas clearly indicate the ways to improve accuracy (increasing the baseline or focal length, using smaller pixels, and improving sub-pixel matching accuracy), pointing the way for the iterative upgrade of the system. At the same time, the patent clearly records these core parameters and the constraints they satisfy, making the technical solution of this invention more complete and rigorous, and forming a solid technical barrier.

[0028] The ring-shaped LED marker module also includes a constant current drive circuit and a microcontroller; The ring-shaped LED marking module includes 12 infrared LEDs, which are connected in a 3-series-4-parallel configuration, meaning that every 3 LEDs are connected in series as a group, and a total of 4 groups are connected in parallel, controlled by an independent constant current drive channel. The microcontroller receives commands from the host computer via the UART interface at a baud rate of 115200bps, supports 256 levels of PWM brightness adjustment and three working modes: constant light, low-frequency flicker, and high-frequency modulation, and the PWM dimming frequency is 10kHz. With this configuration, this embodiment provides a ring-shaped LED marking module including a constant current drive circuit and a microcontroller. Its working principle encompasses two levels: electrical connection and communication control. Electrically, the 12 infrared LEDs are connected in a "3-series, 4-parallel" configuration: each group of 3 LEDs is connected in series to ensure each LED receives a consistent drive current, while the 4 groups in parallel reduce the total drive voltage requirement. Each group is controlled by an independent constant current drive channel, ensuring balanced current and stable light emission, avoiding uneven brightness due to individual LED differences. In terms of communication and control, the onboard microcontroller receives instructions from the host computer via a UART serial interface, supporting 256 levels of PWM brightness adjustment. This allows the system to dynamically adjust the LED's luminous intensity based on the working distance and ambient light intensity to achieve optimal image quality and power consumption balance. Simultaneously, it supports three working modes: constant light, low-frequency flicker, and high-frequency modulation. The high-frequency modulation (10kHz PWM) mode, combined with the camera's specific exposure sequence, can effectively distinguish other infrared interference sources in the environment, significantly improving the system's anti-interference capability in complex lighting environments.

[0029] The surgical navigation module includes: The preoperative image registration unit is used to read and reconstruct 3D CT image data and define the image coordinate system. The spatial registration unit achieves spatial mapping between the patient's actual position and the image coordinate system through point cloud matching or marker point registration algorithms. The instrument pose mapping unit is used to convert the surgical instrument pose output by the image processing module to the image coordinate system in real time. The path planning and early warning unit is used to display the preset surgical path and issue an audible and visual alarm signal when the surgical instrument navigation and tracking is lost; The visualization display unit presents the spatial positional relationship between surgical instruments and the patient's anatomical structure in real time on the display terminal in a multi-view manner; Through the above technical solution, this embodiment provides a surgical navigation module, including a preoperative image registration unit, a spatial registration unit, an instrument pose mapping unit, a path planning warning unit, and a visualization display unit. The preoperative image registration unit is responsible for reading the patient's preoperative CT three-dimensional image data and performing three-dimensional reconstruction to establish a digital model of the patient's anatomical structure and the corresponding image coordinate system. The spatial registration unit is the core component of this module. It uses point cloud matching algorithms (such as the ICP iterative nearest point algorithm) or externally pasted marker point registration algorithms to accurately align the patient's actual spatial position on the operating table with the digital model, establishing a one-to-one mapping relationship from physical space to image space. The instrument pose mapping unit uses the above mapping relationship to convert the real-time pose data of surgical instruments obtained from the image processing module and located in the physical camera coordinate system to the image coordinate system in real time, thereby achieving "coexistence" of instruments and anatomical models in the same virtual space. The path planning warning unit is responsible for displaying the preoperatively planned surgical path and issuing a warning when the system detects tracking failure. The visualization display unit presents the fused scene to the doctor in a visually intuitive way using multiple views (such as 3D perspective view, axial / coronal / sagittal section view); By clearly dividing the functions of the surgical navigation module into distinct units, a complete and efficient processing pipeline from image data to real-time intraoperative guidance is constructed. Each unit has a clear responsibility and works closely together, ensuring the accuracy, real-time nature, and intuitiveness of the navigation information. Through multi-view visualization, doctors can simultaneously obtain three-dimensional spatial perception and two-dimensional sectional details, fully grasping the relative positional relationship between instruments and key anatomical structures (such as blood vessels, nerves, and lesions). This design not only makes the presentation of navigation information more in line with doctors' cognitive habits and improves the intuitiveness of surgical operations, but also provides another layer of protection for surgical safety through the path planning and early warning function. The design of the entire module fully reflects the pragmatic concept of being doctor-centered and clinically demand-oriented.

[0030] In summary, this embodiment provides a surgical navigation system based on optical motion capture and ring LED markers. The ring LED marker module creates an anti-occlusion infrared spot array; the image processing module uses a weighted optimization algorithm to achieve high-precision pose calculation; the data analysis module uses hierarchical judgment and statistical diagnosis to achieve real-time monitoring of navigation accuracy and anomaly tracing; and the surgical navigation module transforms pose data into an intuitive intraoperative navigation view. All modules work collaboratively through clearly defined data interfaces and status codes, forming a complete closed loop from data acquisition to clinical guidance, providing a stable, accurate, and intelligent navigation solution for minimally invasive surgery.

[0031] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A surgical navigation system based on optical motion capture and ring LED marking, characterized in that, The system includes: The ring-shaped LED marking module, fixed to the surgical instrument, includes an LED ring light, which is composed of multiple infrared LEDs evenly distributed on a ring substrate, used to form an array of infrared light spots that can be recognized by a binocular infrared camera; The infrared camera module includes two binocular infrared cameras mounted at a fixed baseline distance for synchronously acquiring images of an infrared spot array; The image processing module is used to receive images of the infrared light spot array, extract the center sub-pixel coordinates of each LED light spot, calculate the visibility score of each LED based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera, assign confidence weights to each LED according to the visibility scores, and then solve the three-dimensional spatial position and orientation of the surgical instrument and the weighted projection error index by minimizing the weighted projection error function. The data analysis module compares the weighted projection error index with a preset error threshold range to classify the tracking status. When the tracking is judged to be abnormal, it further performs statistical analysis on the individual weighted projection error of each visible LED to distinguish between system calibration drift and local anomalies. The surgical navigation module includes a preoperative image registration unit, a spatial registration unit, an instrument pose mapping unit, a path planning and early warning unit, and a visualization display unit; The surgical navigation module is used to establish a three-dimensional spatial reference image coordinate system based on preoperative image data, spatially register the patient's actual position with the preoperative image, map the surgical instrument pose output by the image processing module to the image coordinate system in real time, and display the positional relationship of the surgical instruments relative to the anatomical structure in a visual manner in real time.

2. The surgical navigation system based on optical motion capture and ring LED marking according to claim 1, characterized in that, The processing procedure of the image processing module includes: The three-dimensional spatial position and orientation of the surgical instrument, as well as the weighted projection error exponent, are solved by minimizing the weighted projection error function, as shown in the following formula: ； in, To add a weighted projection error index, 3 3. Rotation matrix, representing the orientation of the surgical instruments relative to the world coordinate system. 3 1. Translation vector, representing the spatial position of the surgical instrument in the world coordinate system; S, the set of visible LED points, which is the set of all LEDs successfully identified by the binocular infrared camera system at the current moment; i, an integer index, representing the number of the visible LED. Let be the confidence weight for the i-th LED, with a value between 0 and 1. Let i be the known three-dimensional coordinates of the i-th LED in the world coordinate system. Let i be the observed three-dimensional coordinates of the i-th LED in the camera coordinate system. Let be the projection transformation function. This is the symbol for the Euclidean norm.

3. A surgical navigation system based on optical motion capture and ring LED marking according to claim 2, characterized in that, The analysis process of the data analysis module includes: By weighted projection error index Compared with the preset error threshold range Perform a comparison; like If the system determines that the tracking is lost, it will automatically freeze the navigation screen, stop outputting pose data to the surgical navigation module, and issue an audible and visual alarm signal. like If the tracking is deemed abnormal, an alert will be issued; like It was judged to be a high-quality tracker.

4. A surgical navigation system based on optical motion capture and ring LED marking according to claim 3, characterized in that, The analysis process of the data analysis module also includes: When an anomaly is identified in the tracking, a more detailed analysis is performed based on the weighted projection error index: S1: Calculate the individual weight projection error index of each visible LED at preset time intervals during the operation; S2: Construct an error distribution vector based on the individual weight projection error index of each visible LED, and calculate the mean, standard deviation, and maximum deviation of the error distribution vector; S3: By comparing the mean, standard deviation, and maximum deviation of all visible LEDs with preset mean threshold, standard deviation threshold, and maximum deviation threshold respectively, and if any parameter is higher than the threshold, it is determined that the system calibration accuracy has exceeded the tolerance, the pose output is automatically paused and a recalibration prompt is sent to the operator.

5. A surgical navigation system based on optical motion capture and ring LED marking according to claim 1, characterized in that, The image processing procedure of the image processing module also includes: First, the visibility score of the i-th LED is calculated based on the angle between the normal direction of the emitting surface of each LED and the optical axis of the camera and a preset angular constant. Then, the confidence weight of the i-th LED is calculated based on the visibility score.

6. A surgical navigation system based on optical motion capture and ring LED marking according to claim 1, characterized in that, The two binocular infrared cameras in the infrared camera module satisfy the following accuracy constraints: The depth positioning error is equal to the square of the target distance divided by the product of the baseline distance and the equivalent focal length, and then multiplied by the parallax matching accuracy. The baseline distance must be no less than the square of the maximum working distance divided by the equivalent focal length, and then multiplied by the ratio of the parallax matching accuracy to the maximum permissible depth positioning error.

7. A surgical navigation system based on optical motion capture and ring LED marking according to claim 1, characterized in that, The ring-shaped LED marker module also includes a constant current drive circuit and a microcontroller; The ring-shaped LED marking module includes 12 infrared LEDs, which are connected in a 3-series-4-parallel configuration, meaning that every 3 LEDs are connected in series as a group, and a total of 4 groups are connected in parallel, controlled by an independent constant current drive channel. The microcontroller receives commands from the host computer via the UART interface at a baud rate of 115200bps, supports 256 levels of PWM brightness adjustment and three working modes: constant light, low-frequency flicker, and high-frequency modulation, and the PWM dimming frequency is 10kHz.

8. A surgical navigation system based on optical motion capture and ring LED marking according to claim 1, characterized in that, The surgical navigation module includes: The preoperative image registration unit is used to read and reconstruct 3D CT image data and define the image coordinate system. The spatial registration unit achieves spatial mapping between the patient's actual position and the image coordinate system through point cloud matching or marker point registration algorithms. The instrument pose mapping unit is used to convert the surgical instrument pose output by the image processing module to the image coordinate system in real time. The path planning and early warning unit is used to display the preset surgical path and issue an audible and visual alarm signal when the surgical instrument navigation and tracking is lost; The visualization display unit presents the spatial positional relationship between surgical instruments and the patient's anatomical structure in real time on the display terminal in a multi-view manner.

Images