Astigmatism lens implantation axis navigation method and system based on double-body position eye images

Abstract

The application discloses a method and system for astigmatism lens implantation axis navigation based on double-body-position eye images and belongs to the technical field of astigmatism lens implantation axis navigation, and comprises the following steps: S1. sitting image acquisition: a mobile phone or a tablet mobile terminal is used to take an image of the eyes of a sitting cataract astigmatism patient; S2. sitting parameter measurement: inherent characteristics of irises and conjunctival vascular textures are used as positioning references; S3. lying image acquisition: fixed mobile terminal shooting parameters are used; S4. lying parameter measurement: the same inherent eye characteristic reference is used, and a corneal limbus is identified; S5. rotation correction navigation: sitting and lying astigmatism axis data are compared, and the corresponding eyeball rotation angle of body position conversion is calculated; the application relies on lifelong stable and unchanged eye anatomical characteristics such as conjunctival vascular inflection points and iris ravine textures, and implements rigid registration on two groups of eye images in the sitting and lying positions, so that the sitting acquisition data are used as anatomical reference to calculate the physiological rotation amount of the eyeball caused by body position switching and to perform reverse compensation.

Description

Technical Field

[0001] This invention relates to the field of axis navigation technology for astigmatic lens implantation, and more specifically, to an axis navigation method and system for astigmatic lens implantation based on dual-position eye images. Background Technology

[0002] In patients with cataracts and corneal astigmatism undergoing phacoemulsification combined with Toric intraocular lens implantation, the accuracy of the astigmatic lens implantation axis directly determines the amount of residual astigmatism, the quality of uncorrected visual acuity recovery, and visual comfort postoperatively. Ideally, the lens should be implanted strictly along the principal diameter of the corneal anatomical astigmatism. Every 1° deviation in axis results in approximately a 3.3% decrease in astigmatism correction efficiency. Deviations exceeding 5° can cause significant double vision, dizziness, and visual fatigue. Deviations of 10° or more will completely eliminate the astigmatism correction effect, requiring a second surgery to adjust the lens position, significantly increasing medical risks and the patient's financial burden. Currently, the mainstream clinical positioning methods are divided into two categories: manual marker positioning and intraoperative OCT navigation positioning. Both have unavoidable technical shortcomings, making it difficult to balance accuracy, cost, patient suitability, and accessibility at the grassroots level.

[0003] Therefore, we have made improvements to this and proposed an axis navigation method and system for astigmatic lens implantation based on dual-position eye images. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, the purpose of this invention is to provide a method and system for axial navigation of astigmatic lens implantation based on dual-position eye images.

[0005] To solve the above problems, the technical solution adopted by the present invention is as follows: the following steps: S1. Seated Image Acquisition: Using a mobile phone or tablet, images are captured of the eyes of a seated cataract astigmatism patient. No manual marking with a slit lamp is required, making it suitable for elderly patients with poor cooperation. S2. Seated parameter calculation: Using the inherent features of the vascular texture of the iris and conjunctiva as the positioning reference, the corneal limbus contour is identified, and the corneal center coordinates and seated astigmatic axis data are calculated. S3. Supine Image Acquisition: Fix the shooting parameters of the mobile terminal, switch the patient to the surgical supine position, and shoot the same eye to obtain supine eye images; S4. Calculation of parameters in supine position: Using the same inherent characteristics of the eye as a benchmark, the limbus is identified, and the coordinates of the corneal center and the astigmatic axis data in supine position are calculated. S5. Rotational Correction Navigation: By comparing astigmatism axis data in sitting and lying positions, the eye rotation angle corresponding to the change in body position is calculated, and the initial axis is corrected to obtain the precise implantation axis during surgery. Low-cost digital navigation can be completed using a mobile device, replacing the intraoperative OCT equipment.

[0006] Preferably, during the image acquisition process in sitting and supine positions, the imaging acquisition system includes a mobile imaging device, a fixed imaging parameter calibration module, and a shooting posture calibration module. The mobile imaging device locks and fixes the focal length, shooting distance, and imaging resolution, with the shooting distance calibrated to the standard near-field imaging range of 8-12cm. The shooting posture calibration module ensures that the device lens is always perpendicular to the geometric tangent plane of the patient's operated cornea for forward shooting. The fixed imaging parameter calibration module unifies the pixel physical magnification, corneal imaging size, field of view coverage, and scaling ratio of the two images in sitting and supine positions, eliminating framing distortion and perspective angle deviation.

[0007] Preferably, during the image parameter calculation and image registration process, the feature matching and calibration system stores a database of inherent stable anatomical features of the operated eye, specifically including inflection points of superficial conjunctival vessels, nodes of gradual change in vessel thickness, vessel trajectory, iris groove texture, and iris pigment distribution features; the feature matching and calibration system selects no less than 8 sets of evenly distributed, highly recognizable feature points, and completes point-to-point rigid registration and alignment of the dual-position images through feature point spacing comparison and matching.

[0008] Preferably, the corneal parameter precision measurement system is equipped with a contour filtering module, a corneal center solving module, and an astigmatic axis fitting module. The contour filtering module identifies pixels point by point across the entire domain, eliminating invalid contour points caused by eyelashes, eyelids, tear reflections, and conjunctival folds, and fitting a complete and closed corneal limbus ring effective contour. The corneal center solving module calculates the corneal geometric centroid through a weighted average calculation of the entire domain pixel coordinates and defines it as the standard corneal center coordinates. The astigmatic axis fitting module establishes an ocular polar coordinate system with the corneal center as the origin, combining the corneal refractive power distribution and ellipsoidal structure.

[0009] Preferably, the eye rotation error calibration system has a built-in body position difference comparison model, which uses the astigmatic axis calculated by sitting imaging as the anatomical reference axis and matches the axis data calculated by supine imaging. The system presets a human physiological eye rotation offset threshold of ≤10°, and accurately calculates the eye rotation offset caused by body position changes by comparing the difference between the two body position axis data.

[0010] The preferred low-cost precision navigation alternative system abandons the traditional slit lamp manual marking structure and high-cost intraoperative OCT navigation hardware, and relies on general mobile devices to complete preoperative axis calibration; the system avoids the problems of marking failure caused by subjective human interpretation, marking deviation and poor patient cooperation, and compresses the marking error of traditional manual ±3°~5° to within ±0.5°.

[0011] Preferably, the high-definition anti-interference imaging supplementary light system uses uniform diffuse reflection natural light or low-brightness soft cold light source to precisely control the imaging illumination intensity at 300-500lx; the system avoids corneal reflection and eyelid and eyelash projection obstruction caused by strong direct light and point light source irradiation, and completely preserves the iris texture, pigment distribution and conjunctival vascular structure of the operated eye, so that the effective feature point preservation rate is ≥95%.

[0012] Preferably, the multi-feature cross-validation registration system can uniformly select 10-20 sets of conjunctival blood vessel and iris texture feature points distributed throughout the imaging area of ​​the operated eye to construct a patient-specific feature matching system; the system is equipped with a multi-feature point similarity weighted matching model to achieve multi-point cross-validation matching and avoid matching failure caused by single feature occlusion or blurring.

[0013] Preferably, the system uses a smartphone or tablet as the sole hardware carrier, equipped with a CMOS image sensor and an embedded processing chip, eliminating the need for intraoperative OCT equipment and slit lamp manual marking fixtures; specifically, it includes the following modules: The system integrates a high-definition anti-interference supplementary lighting module, an imaging acquisition and calibration module, a feature matching and calibration module, a multi-feature cross-validation registration module, a corneal parameter calculation module, an eye rotation error calibration module, and a low-cost navigation and computing module. These modules work in synergy, using serial data interaction and a closed-loop verification mechanism to complete the entire process of image acquisition, preprocessing, feature registration, parameter calculation, error correction, and intraoperative axis output. The high-definition anti-interference supplementary lighting module uses photosensitivity detection and closed-loop dimming to stabilize the imaging illumination within the 300-500 lx diffuse soft light range, avoiding corneal reflection and eyelid / eyelash obstruction, thus preserving the complete anatomical features of the eye. The imaging acquisition and calibration module fixes the imaging focal length, resolution, and 8-12 cm standard imaging object distance, combined with the design... The gyroscope is used to calibrate the shooting posture, ensuring that the lens is perpendicular to the corneal section of the operated eye, and to unify the pixel magnification, scaling scale, and field of view of the dual-position images. The feature matching and calibration module stores stable anatomical features such as conjunctival vessels and iris texture, and selects at least 8 sets of uniform and highly recognizable feature points as registration benchmarks through grayscale thresholding. The multi-feature cross-validation registration module relies on a grayscale and structural similarity weighted algorithm to cross-validate 10-20 sets of global feature points, eliminate abnormal points, and complete the coordinate system of the dual-position images. The corneal parameter calculation module filters out invalid pixels such as eyelashes, tears, and conjunctival folds, and fits the complete corneal limbus closure contour. The eye rotation error calibration module uses the seated axis as the anatomical benchmark, combined with a physiological rotation threshold of ≤10°.

[0014] Preferably, the system adopts a hierarchical closed-loop architecture of optical calibration, standardized imaging, multi-feature registration, parameter solving, error correction, and result output. Each module relies on embedded fixed algorithms to achieve high-precision automated operation. The high-definition anti-interference supplementary lighting module, combined with a diffuse reflection soft light structure and brightness closed-loop adjustment, stabilizes the preoperative imaging lighting environment, ensuring an effective feature point retention rate of ≥95%, providing high-quality original images for image registration. The imaging acquisition and calibration module, through parameter solidification and posture sensing calibration, achieves complete normalization of dual-position imaging parameters, eliminating systematic errors caused by differences in shooting posture, object distance, and magnification, ensuring accurate comparison of the two sets of images. The multi-feature cross-validation registration module integrates feature grayscale, texture morphology, and spatial position. By incorporating multidimensional information and employing weighted similarity joint verification to eliminate feature drift points, subpixel-level image registration is achieved, ensuring precise alignment of the spatial positions of the operated eyes in both body positions. The corneal parameter calculation module uses adaptive filtering for noise reduction, accurately fitting the corneal limbus contour, and offsetting local deviations by using a weighted average of pixels across the entire domain. It also combines the polar coordinate system of the eye with the refractive characteristics of the cornea to solve for a high-precision astigmatic axis. The eye rotation error calibration module models the physiological movement characteristics of the eye, distinguishes between physiological rotation deviation and imaging noise error, accurately calculates the eye rotation angle during body position switching, and dynamically corrects the axis. The system operates fully automatically, eliminating the need for manual marking and subjective interpretation, compressing the error of traditional manual marking (±3°~5°) to within ±0.5°.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention relies on lifelong, stable ocular anatomical features such as conjunctival vascular inflection points and iris groove textures to perform rigid registration on two sets of ocular images in sitting and supine positions. Using data acquired in the sitting position as an anatomical benchmark, it calculates and compensates for the physiological rotation of the eyeball caused by positional changes, reducing the original 3°–5° positioning error of manual marking to ±0.5°. Local axis fitting accuracy can reach ±0.25°, significantly improving the correction effect. This method uniformly distributes multiple sets of feature points within the surgical eye's visual field for cross-validation and matching, even in cases where the image is obscured by eyelashes, tear reflection, etc. Even in cases of localized image blurring, feature alignment can still be stably completed, with registration errors consistently controlled within ±0.2°. The system automatically removes invalid contour pixels caused by eyelid and conjunctival folds and specular reflections, reconstructing a complete and continuous limbal contour. The geometric center of the cornea is determined through a weighted average algorithm, and the principal and secondary astigmatic diameter angles are accurately fitted using the polar coordinate system of the eye. This effectively eliminates calculation deviations caused by body rotation, shooting distortion, and local occlusion, reducing postoperative discomfort such as residual astigmatism, double vision, and eye strain, and lowering the surgical risk of needing to readjust the lens position due to positioning deviations. The entire process eliminates the need for manual corneal surface marking with a slit lamp. Combined with soft illumination from a 300-500 lx diffuse cold light source, it avoids causing photophobia and eye closure due to strong light stimulation. It also eliminates the need for patients to keep their head and eyes fixed for extended periods. Patients with low cooperation levels, such as the elderly, those with cognitive impairments, or those with limb stiffness, can successfully complete the imaging process. The effective anatomical feature retention rate of the operated eye remains stable at over 95%, solving the clinical challenge of preoperative marking for special patients and broadening the applicability of precise astigmatic lens implantation surgery. Attached Figure Description

[0016] Figure 1 A flowchart of the astigmatic lens implantation axis navigation method and system based on dual-position eye images provided in this application. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0018] A method and system for axial navigation of astigmatic lens implantation based on dual-position ocular images, comprising the following steps: S1. Seated Image Acquisition: Using a mobile phone or tablet, images are captured of the eyes of a seated cataract astigmatism patient. No manual marking with a slit lamp is required, making it suitable for elderly patients with poor cooperation. S2. Seated parameter calculation: Using the inherent features of the vascular texture of the iris and conjunctiva as the positioning reference, the corneal limbus contour is identified, and the corneal center coordinates and seated astigmatic axis data are calculated. S3. Supine Image Acquisition: Fix the shooting parameters of the mobile terminal, switch the patient to the surgical supine position, and shoot the same eye to obtain supine eye images; S4. Calculation of parameters in supine position: Using the same inherent characteristics of the eye as a benchmark, the limbus is identified, and the coordinates of the corneal center and the astigmatic axis data in supine position are calculated. S5. Rotational Correction Navigation: By comparing astigmatism axis data in sitting and lying positions, the eye rotation angle corresponding to the change in body position is calculated, and the initial axis is corrected to obtain the precise implantation axis during surgery. Low-cost digital navigation can be completed using a mobile device, replacing the intraoperative OCT equipment.

[0019] Furthermore, during the image acquisition process in sitting and lying positions, the imaging system includes a mobile imaging device, a fixed imaging parameter calibration module, and a shooting posture calibration module. The mobile imaging device locks and fixes the focal length, shooting distance, and imaging resolution, with the shooting distance calibrated to the standard near-field imaging range of 8-12cm. The shooting posture calibration module ensures that the device lens is always perpendicular to the geometric tangent plane of the patient's operated cornea for forward shooting. The fixed imaging parameter calibration module unifies the pixel physical magnification, corneal imaging size, field of view coverage, and scaling ratio of the two images in sitting and lying positions, eliminating framing distortion and perspective angle deviation, and controlling the overall imaging matching accuracy within ±0.5°.

[0020] Furthermore, during image parameter calculation and image registration, the feature matching and calibration system stores a database of inherent stable anatomical features of the operated eye, specifically including inflection points of superficial conjunctival vessels, nodes of gradual changes in vessel thickness, vessel trajectories, iris groove texture, and iris pigment distribution characteristics. The feature matching and calibration system selects no fewer than 8 sets of evenly distributed, highly recognizable feature points, and completes point-to-point rigid registration and alignment of dual-position images through feature point spacing comparison and matching. Relying on the characteristics of human anatomical features that remain unchanged throughout life and are not affected by body position or eyeball micro-movements, it eliminates matching misalignment and axis measurement errors caused by image translation and local deflection.

[0021] Furthermore, the corneal parameter precision calculation system is equipped with a contour filtering module, a corneal center solving module, and an astigmatic axis fitting module. The contour filtering module identifies pixels point by point across the entire domain, eliminating invalid contour points caused by eyelashes, eyelids, tear reflections, and conjunctival folds, and fitting a complete and closed corneal limbus ring effective contour. The corneal center solving module calculates the corneal geometric centroid using a weighted average calculation of the entire domain pixel coordinates and defines it as the standard corneal center coordinates. The astigmatic axis fitting module establishes an ocular polar coordinate system with the corneal center as the origin, and combines the corneal refractive power distribution and ellipsoidal structure to accurately fit the principal and secondary astigmatic diameter angles through contour fitting, with an axis calculation accuracy of ±0.25°.

[0022] Furthermore, the eye rotation error calibration system incorporates a positional difference comparison model, using the astigmatic axis calculated by seated imaging as the anatomical reference axis and matching the axis data calculated by supine imaging. The system presets a human physiological eye rotation offset threshold of ≤10°. Through the comparison of axis data differences between the two positions, it accurately calculates the eye rotation offset caused by positional changes. Based on the offset, it performs reverse angle compensation on the supine deviation axis, eliminating the positioning error caused by positional eye rotation and restoring the true anatomical astigmatic axis of the cornea.

[0023] Furthermore, the low-cost precision navigation replacement system abandons the traditional slit-lamp manual marking structure and high-cost intraoperative OCT navigation hardware, relying on general mobile devices to complete preoperative axis calibration; the system avoids the problems of marking failure caused by subjective human interpretation, marking deviation, and poor patient cooperation, compressing the marking error of traditional manual ±3°~5° to within ±0.5°, making it suitable for elderly patients, patients with cognitive impairment, patients with limb stiffness and low cooperation, and suitable for widespread use in hospitals at all levels.

[0024] Furthermore, the high-definition anti-interference imaging supplementary lighting system uses uniform diffuse natural light or low-brightness soft cold light source to precisely control the imaging illumination intensity between 300-500 lx. The system avoids corneal reflection and eyelid and eyelash projection occlusion caused by strong direct light and point light source illumination, and completely preserves the iris texture, pigment distribution and conjunctival vascular structure of the operated eye, so that the effective feature point retention rate is ≥95%, providing distortion-free and high-fidelity original image data support for image registration, contour fitting and axis calculation.

[0025] Furthermore, the multi-feature cross-validation registration system can uniformly select 10-20 sets of globally distributed conjunctival blood vessel and iris texture feature points within the imaging area of ​​the operated eye to construct a patient-specific feature matching system. The system is equipped with a multi-feature point similarity weighted matching model to achieve multi-point cross-validation matching, avoiding matching failure caused by single feature occlusion or blurring, controlling the image registration error within ±0.2°, and ensuring the authenticity and accuracy of the eye rotation angle calculation.

[0026] Furthermore, the system uses a smartphone or tablet as its sole hardware platform, equipped with a CMOS image sensor and embedded processing chip, eliminating the need for intraoperative OCT equipment and slit lamp manual marking fixtures; specifically, it includes the following modules: The system integrates a high-definition anti-interference supplementary lighting module, an imaging acquisition and calibration module, a feature matching and calibration module, a multi-feature cross-validation and registration module, a corneal parameter calculation module, an eye rotation error calibration module, and a low-cost navigation and computing module. These modules work in synergy, using serial data interaction and a closed-loop verification mechanism to complete the entire process of image acquisition, preprocessing, feature registration, parameter calculation, error correction, and intraoperative axis output. The high-definition anti-interference supplementary lighting module uses photosensitivity detection and closed-loop dimming to stabilize the imaging illumination within the 300-500 lx diffuse soft light range, avoiding corneal reflection and eyelid / eyelash obstruction, thus preserving the complete anatomical features of the eye. The imaging acquisition and calibration module fixes the imaging focal length, resolution, and 8-12 cm standard imaging object distance, and, combined with the device's gyroscope, calibrates the shooting posture to ensure the lens is perpendicular to the corneal section of the operated eye, unifying the pixel magnification, scaling, and field of view of dual-position images. The feature matching and calibration module... The registration module stores stable anatomical features such as conjunctival vessels and iris texture, and selects at least 8 sets of uniform, high-discrimination feature points as registration benchmarks through grayscale thresholding. The multi-feature cross-validation registration module relies on a grayscale and structural similarity weighted algorithm to cross-validate 10-20 sets of global feature points, eliminate abnormal points, and complete the dual-position image coordinate system. The corneal parameter calculation module filters out invalid pixels such as eyelashes, tears, and conjunctival folds, fits the complete corneal limbus closure contour, solves the corneal geometric center through pixel weighted averaging, and fits the astigmatic axis in combination with the corneal ellipsoidal refractive structure. The eye rotation error calibration module uses the seated axis as the anatomical benchmark, combines a physiological rotation threshold of ≤10°, and accurately calculates and compensates for eye rotation offset through a positional difference comparison model. The low-cost navigation calculation module performs coordinate transformation on the corrected axis data and outputs the standard navigation axis for intraoperative lens implantation.

[0027] Furthermore, the system adopts a hierarchical closed-loop architecture of optical calibration, standardized imaging, multi-feature registration, parameter solving, error correction, and result output. Each module relies on embedded fixed algorithms to achieve high-precision automated operation. The high-definition anti-interference supplementary lighting module, combined with a diffuse reflection soft light structure and brightness closed-loop adjustment, stabilizes the preoperative imaging lighting environment, ensuring an effective feature point retention rate of ≥95%, providing high-quality original images for image registration. The imaging acquisition and calibration module, through parameter solidification and posture sensing calibration, achieves complete normalization of dual-position imaging parameters, eliminating system errors caused by differences in shooting posture, object distance, and magnification, ensuring accurate comparison of the two sets of images. The multi-feature cross-validation registration module integrates multi-dimensional information such as feature grayscale, texture morphology, and spatial position, and through weighted similarity joint verification, eliminates feature drift points, achieving... The system employs subpixel-level image registration to ensure precise alignment of the surgical eyes in both patient positions. The corneal parameter calculation module utilizes adaptive filtering and noise reduction to accurately fit the corneal limbus contour. It compensates for local deviations using a global pixel-weighted average and combines the ocular polar coordinate system with corneal refractive characteristics to solve for high-precision astigmatic axis alignment. The eye rotation error calibration module models the physiological movement characteristics of the eye, distinguishing between physiological rotation deviation and imaging noise error. It accurately calculates the eye rotation angle during position switching and dynamically corrects the axis alignment. The system operates fully automatically, eliminating the need for manual marking and subjective interpretation. It reduces the error of traditional manual marking (±3°–5°) to within ±0.5°, replacing high-cost intraoperative OCT hardware. Relying on a universal mobile terminal, it enables digital navigation of the astigmatic lens implantation axis in a universal, low-cost, high-precision, and highly interference-resistant manner within the operating room.

[0028] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of protection of this application is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of one or more embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0029] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

Claims

1. A method for axis navigation in astigmatic lens implantation based on dual-position eye images, characterized in that, Includes the following steps: S1. Seated Image Acquisition: Using a mobile phone or tablet, images are captured of the eyes of a seated cataract astigmatism patient. S2. Seated parameter calculation: Using the inherent features of the vascular texture of the iris and conjunctiva as the positioning reference, the corneal limbus contour is identified, and the corneal center coordinates and seated astigmatic axis data are calculated. S3. Supine Image Acquisition: Fix the shooting parameters of the mobile terminal, switch the patient to the surgical supine position, and shoot the same eye; S4. Calculation of parameters in supine position: Using the same inherent characteristics of the eye as a benchmark, the limbus is identified, and the coordinates of the corneal center and the astigmatic axis data in supine position are calculated. S5. Rotational Correction Navigation: By comparing astigmatism axis data in sitting and lying positions, the eye rotation angle corresponding to the change in body position is calculated, and the initial axis is corrected to obtain the precise implantation axis during surgery. Low-cost digital navigation can be completed using a mobile device, replacing the intraoperative OCT equipment.

2. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, During the image acquisition process in sitting and supine positions, the imaging system includes a mobile imaging device, a fixed imaging parameter calibration module, and a shooting posture calibration module. The mobile imaging device locks a fixed focal length, a fixed shooting distance, and a fixed imaging resolution, with the shooting distance calibrated to the standard near-field imaging range of 8-12cm. The shooting posture calibration module ensures that the device lens is always perpendicular to the geometric tangent plane of the patient's operated cornea for forward shooting. The fixed imaging parameter calibration module standardizes the pixel physical magnification, corneal imaging size, field of view coverage, and scaling ratio of the two images in sitting and supine positions.

3. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, During image parameter calculation and image registration, the feature matching and calibration system stores a database of inherent stable anatomical features of the operated eye, specifically including inflection points of superficial conjunctival vessels, nodes of gradual change in vessel thickness, vessel trajectory, iris groove texture, and iris pigment distribution features; the feature matching and calibration system selects no less than 8 sets of evenly distributed and highly recognizable feature points.

4. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, The corneal parameter precision measurement system is equipped with a contour filtering module, a corneal center solving module, and an astigmatic axis fitting module. The contour filtering module identifies pixels point by point across the entire domain, eliminating invalid contour points caused by eyelashes, eyelids, tear reflections, and conjunctival folds, and fitting a complete and closed corneal limbus ring effective contour. The corneal center solving module calculates the corneal geometric centroid through a weighted average calculation of the entire domain pixel coordinates and defines it as the standard corneal center coordinates. The astigmatic axis fitting module establishes an ocular polar coordinate system with the corneal center as the origin, combining the corneal refractive power distribution and ellipsoidal structure.

5. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, The eye rotation error calibration system has a built-in body position difference comparison model. It uses the astigmatic axis calculated by sitting imaging as the anatomical reference axis and matches the axis data calculated by supine imaging. The system presets a human physiological eye rotation offset threshold of ≤10°. Through the comparison of the axis data difference between the two body positions, it accurately calculates the eye rotation offset caused by the change in body position.

6. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, The low-cost precision navigation replacement system abandons the traditional slit lamp manual marking structure and high-cost intraoperative OCT navigation hardware, and relies on general mobile devices to complete preoperative axis calibration. The system avoids the problems of marking failure caused by subjective human interpretation, marking deviation and poor patient cooperation, and compresses the marking error of traditional manual ±3°~5° to within ±0.5°.

7. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, The high-definition anti-interference imaging supplementary light system uses uniform diffuse reflection natural light or low-brightness soft cold light source to precisely control the imaging illumination intensity at 300-500lx. The system avoids corneal reflection and eyelid and eyelash projection obstruction caused by strong direct light and point light source irradiation, and completely preserves the iris texture, pigment distribution and conjunctival vascular structure of the operated eye, so that the effective feature point preservation rate is ≥95%.

8. The method for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 1, characterized in that, The multi-feature cross-validation registration system can uniformly select 10-20 sets of globally distributed conjunctival blood vessel and iris texture feature points within the imaging area of ​​the operated eye to construct a patient-specific feature matching system; the system is equipped with a multi-feature point similarity weighted matching model.

9. A system for axis navigation of astigmatic lens implantation based on dual-position ocular images, the method for axis navigation of astigmatic lens implantation based on dual-position ocular images according to any one of claims 1-8, characterized in that, Specifically, it includes the following modules: The system includes a high-definition anti-interference supplementary lighting module, an imaging acquisition and calibration module, a feature matching and calibration module, a multi-feature cross-validation and registration module, a corneal parameter calculation module, an eye rotation error calibration module, and a low-cost navigation and computing module. The high-definition anti-interference supplementary lighting module uses photosensitivity detection and closed-loop dimming to stabilize the imaging illumination within the 300-500 lx diffuse soft light range, fully preserving the anatomical features of the eye. The imaging acquisition and calibration module fixes the imaging focal length, resolution, and standard imaging object distance of 8-12 cm, and, combined with the device's gyroscope, calibrates the shooting posture to ensure the lens is perpendicular to the corneal section of the operated eye, unifying the dual-position image. The system includes a magnification scale, scaling scale, and field of view. The feature matching and calibration module stores stable anatomical features such as conjunctival vessels and iris texture, and selects at least 8 sets of uniform, high-discrimination feature points as registration benchmarks through grayscale thresholding. The multi-feature cross-validation registration module relies on a grayscale and structural similarity weighted algorithm to cross-validate 10-20 sets of global feature points and eliminate abnormal points. The corneal parameter calculation module filters out invalid pixels such as eyelashes, tears, and conjunctival folds and fits a complete corneal limbal closure contour. The eye rotation error calibration module uses the seated axis as the anatomical benchmark and combines a physiological rotation threshold of ≤10°.

10. The method and system for axis navigation of astigmatic lens implantation based on dual-position eye images according to claim 9, characterized in that, The high-definition anti-interference supplementary lighting module combines a diffuse reflection soft light structure with closed-loop brightness adjustment; the imaging acquisition and calibration module achieves complete normalization of dual-position imaging parameters through parameter solidification and posture sensing calibration, eliminating systematic errors caused by differences in shooting posture, object distance, and magnification, and ensuring accurate comparison of the two sets of images; the multi-feature cross-verification registration module integrates multi-dimensional information such as feature grayscale, texture morphology, and spatial position, and eliminates feature drift points through weighted similarity joint verification, achieving sub-pixel-level image registration and ensuring accurate alignment of the two-position operated eyes in spatial position; the corneal parameter calculation module adopts adaptive filtering noise reduction, accurately fits the corneal limbal contour, relies on global pixel weighted averaging to offset local deviations, and combines the ocular polar coordinate system and corneal refractive characteristics to solve for high-precision astigmatic axis; the eye rotation error calibration module models based on the physiological movement characteristics of the eye, distinguishes between physiological rotation deviation and imaging noise error, accurately calculates the eye rotation angle during position switching, and dynamically corrects the axis.

Images