Control system and control method of cornea crosslinking equipment

Abstract

The invention relates to the technical field of medical equipment control, in particular to a control system and a control method of cornea crosslinking equipment. The control system comprises a multi-modal data fusion module which is configured to integrate a preoperative corneal topographic map curvature matrix, OCT layering thickness data and biomechanical parameters to generate a four-channel fusion tensor; the virtual cornea modeling module is configured to input a conditional generative adversarial network to generate a virtual cornea model under the condition of the four-channel fusion tensor; the dynamic energy planning module is configured to generate a thermodynamic diagram of energy distribution based on mechanical response prediction of the virtual cornea model; and an illumination control module configured to perform patterned illumination based on the thermodynamic diagram. According to the method, the virtual cornea model is constructed, the energy distribution thermodynamic diagram is generated, accurate patterning reference can be provided for irradiation treatment, and a better treatment effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical equipment control, and in particular to a control system and a control method for corneal cross-linking equipment. Background Art

[0002] With the increasing number of myopic patients, corneal cross-linking (CCL) has gained increasing attention as an effective treatment option. Corneal cross-linking (CCL) is a key technology for treating corneal ectatic diseases such as keratoconus. Its core principle is to activate riboflavin through ultraviolet (UVA) irradiation, inducing cross-linking of corneal collagen fibers to enhance mechanical strength. CCL is primarily performed using a corneal cross-linking device. CXL devices are medical devices used to treat corneal diseases such as keratoconus. Their core function is to photochemically enhance the mechanical strength of corneal collagen fibers. However, existing technologies have the following limitations. For example, they lack parameter matching: Traditional cross-linking devices use linear parameters (e.g., a fixed UVA energy / time ratio), which cannot adapt to the complex morphology of keratoconus, such as its asymmetric steep areas and uneven thickness distribution. Risk control is also lagging: Existing systems rely on intraoperative OCT monitoring, which cannot predict the biomechanical response under different energy distributions and carries the risk of postoperative corneal ectasia or perforation (clinical data show that the perforation rate of traditional methods exceeds 3%). Single data utilization: Preoperative examination data (such as Corvis ST biomechanical parameters) are only used for contraindication screening and are not dynamically associated with irradiation parameters.

[0003] Based on this, the existing technology still needs to be improved. Summary of the Invention

[0004] In view of this, the present application proposes a control system and control method for a corneal cross-linking device, the main purpose of which is to overcome the defects of the above-mentioned background technology. By constructing a virtual corneal model and generating a thermal map of energy distribution, an accurate patterned reference is provided for the irradiation treatment of the corneal cross-linking device to achieve better treatment effects.

[0005] A first aspect of the present invention provides a control system for a corneal cross-linking device, comprising a multimodal data fusion module, a virtual corneal modeling module, a dynamic energy planning module, and an irradiation control module. The multimodal data fusion module is configured to integrate the preoperative corneal topography curvature matrix, OCT layered thickness data, and biomechanical parameters to generate a four-channel fusion tensor. The virtual corneal modeling module is configured to generate a virtual corneal model by inputting a conditional generative adversarial network based on the four-channel fusion tensor. The dynamic energy planning module is configured to generate a heat map of energy distribution based on the mechanical response prediction of the virtual corneal model. The irradiation control module is configured to perform patterned irradiation based on the heat map.

[0006] In some embodiments, the multimodal data fusion module includes a curvature processing unit, an OCT analysis unit, a biomechanical mapping unit, and a feature fusion unit. The curvature processing unit is configured to perform Gaussian filtering and normalization on the curvature matrix of the preoperative corneal topography to generate a standardized curvature tensor. The OCT analysis unit is configured to segment the corneal epithelium, stroma, and endothelium in the OCT volume data through a pre-trained neural network, calculate the stroma thickness distribution map, and align it to the coordinate system of the curvature tensor. The biomechanical mapping unit converts the biomechanical parameters obtained by the detection into a biomechanical constraint matrix, and the biomechanical parameters include at least the first applanation deformation amplitude and the curvature radius of the most concave point of the cornea. The feature fusion unit is configured to splice the curvature tensor, the stroma thickness distribution map, and the biomechanical constraint matrix along the channel dimension to form a four-channel fusion tensor, and perform standardization on each channel.

[0007] In some embodiments, the feature fusion unit also includes a risk area enhancer module, which is configured to: identify qualified pixel areas based on a preset normalized thickness threshold, a normalized curvature threshold, and a first flattening deformation amplitude mapping threshold; and multiply the fusion tensor eigenvalue corresponding to the pixel area by an enhancement coefficient.

[0008] In some embodiments, the dynamic energy planning module is configured to: apply an energy attenuation coefficient to an area that meets a stromal layer thickness threshold in a virtual corneal model; generate a light spot arrangement pattern, wherein the diameter of the light spot is 50 to 100 μm, and the light spot density in the central area is 1.5 to 2 times that of the edge area.

[0009] In some embodiments, the control system of the corneal cross-linking device further includes a dynamic optimization module configured to: monitor riboflavin fluorescence intensity in real time; and update the irradiation strategy when the local riboflavin fluorescence intensity is lower than the average of the entire field area by more than a preset ratio.

[0010] In some embodiments, the control system of the corneal cross-linking device also includes a motion compensation module, which is configured to: track pupil edge feature points and calculate the eye deflection vector; establish an eye-robotic arm motion model based on the eye deflection vector; drive the lifting track and the robotic arm to complete synchronous offset within a preset time based on the eye-robotic arm motion model; and make the projection position error of the light spot ≤50μm through closed-loop control.

[0011] In some embodiments, the control system of the corneal cross-linking device further includes a three-level safety fuse module, which includes: a primary protection unit, a secondary protection unit, and a tertiary protection unit. The primary protection unit is configured to detect UVA light density fluctuations in real time and trigger automatic calibration when the fluctuation amplitude exceeds a fluctuation threshold. The secondary protection unit is configured to cut off the UVA light source and lock the robotic arm travel when the distance sensor detects that the probe-cornea distance is less than a distance threshold. The tertiary protection unit is configured to automatically activate an energy attenuation mode for the local area, with an attenuation coefficient of 0.3 to 0.5, when intraoperative OCT detects a localized decrease in stromal layer thickness greater than a preset ratio.

[0012] In some embodiments, the control system of the corneal cross-linking device also includes an intraoperative adjustment module, which is configured to: obtain the actual stromal layer thickness distribution detected by intraoperative OCT at every preset time interval, and adjust the applied energy when the deviation between the actual stromal layer thickness and the predicted stromal layer thickness of the virtual corneal model exceeds a threshold.

[0013] In some embodiments, the control system of the corneal cross-linking device also includes a postoperative evaluation module, which is configured to: compare the changes in corneal biomechanical index before and after surgery; and generate a secondary treatment recommendation report when the corneal biomechanical index decreases to less than an index threshold after a preset time after surgery.

[0014] The second aspect of the present invention also provides a control method for a corneal cross-linking device, which is implemented by the control system described in any of the above embodiments, and includes the following steps: integrating the preoperative corneal topography curvature matrix, OCT layered thickness data and biomechanical parameters to generate a four-channel fusion tensor; using the four-channel fusion tensor as a condition, inputting a conditional generative adversarial network to generate a virtual corneal model; generating a thermal map of energy distribution based on the mechanical response prediction of the virtual corneal model; and performing patterned irradiation based on the thermal map.

[0015] The beneficial effects of the present invention are as follows: Through multimodal data fusion and virtual corneal modeling, this application can generate a highly accurate virtual corneal model, providing a precise basis for irradiation treatment using corneal cross-linking devices. In particular, enhanced modeling of the steep area of ​​the inferior temporal quadrant, where keratoconus is most prevalent, can effectively improve the targeted treatment and success rate. By generating a heat map of energy distribution, it provides a precise reference for irradiation, ensuring the accuracy of energy distribution, effectively avoiding over-irradiation or under-irradiation, and improving the safety and effectiveness of treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other embodiments can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A control system module diagram of a corneal cross-linking device according to one embodiment of the present invention; Figure 2 A flowchart of multimodal data fusion proposed in one embodiment of the present invention; Figure 3 A flowchart for generating a virtual cornea model according to an embodiment of the present invention; Figure 4 A block diagram of the motion compensation control of a robotic arm proposed in one embodiment of the present invention; Figure 5 This is a flow chart of a control method for a corneal cross-linking device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0019] The first aspect of the present invention provides a control system for a corneal cross-linking device. The control system adopts a distributed architecture, with a host computer system deployed on a Windows tablet computer and a slave computer system integrated into the control unit of the treatment device working together. The host computer system is responsible for multi-dimensional fusion analysis, constructing a virtual corneal model through a conditional generative adversarial network, and generating a UVA energy distribution heat map. It also provides a touch-sensitive interactive interface for treatment parameter setting, data visualization, and encrypted treatment record management. The slave computer system is equipped with an irradiation control module, which receives instructions from the host computer via the CAN bus, accurately controls the irradiation intensity and irradiation time of the UVA light source in real time, and simultaneously feeds back the actual output data. The system adopts a dual-redundant safety mechanism, automatically maintaining safe power when communication is interrupted, ensuring the safety and reliability of the treatment process. This architecture fully utilizes the advantages of the host computer in complex calculations and visualization, while using the slave computer to achieve microsecond-level real-time control, meeting the stringent requirements of corneal cross-linking treatment for accuracy and stability.

[0020] Specifically, if Figure 1As shown, the control system of the corneal cross-linking device includes a multimodal data fusion module, a virtual cornea modeling module, a dynamic energy planning module, and an irradiation control module. The multimodal data fusion module is configured to integrate the preoperative corneal topography curvature matrix, OCT layered thickness data, and biomechanical parameters to generate a four-channel fusion tensor (channel dimensions: topographic curvature, thickness gradient, biomechanical strength, lesion probability mask, where the lesion probability mask is derived data generated through collaborative analysis of the first three channels and is used to enhance the weight of the lesion area). The virtual cornea modeling module is configured to use a conditional generative adversarial network (cGAN) architecture, using the four-channel fusion tensor as conditional input to generate a virtual cornea model with enhanced features of the lesion area (particularly the steep area in the inferior temporal quadrant). The dynamic energy planning module is configured to predict the mechanical response of the virtual corneal model and perform voxel-level predictions using a 3D CNN decoder. This generates a heat map of the UVA energy distribution with a spatial resolution of at least 50 μm × 50 μm × 10 μm. Prioritization is given to the steep inferotemporal quadrant, a high-risk area for keratoconus. The irradiation control module is configured to execute patterned irradiation based on the heat map using a digital micromirror device (DMD). The control logic converts the heat map into a DMD bitmap (using the Floyd-Steinberg dithering algorithm) and updates the irradiation pattern every 400 ms.

[0021] Through multimodal data fusion and virtual corneal modeling, this application's control system can generate a highly accurate virtual corneal model. This modeling, in particular, enhances the steep area of ​​the inferotemporal quadrant, where keratoconus is most prevalent. This provides a precise basis for treatment, improving lesion localization accuracy by 40%, effectively enhancing the targeted nature and success rate of treatment. By generating high-spatial-resolution thermal maps, the accuracy of energy distribution is ensured, effectively avoiding over- or under-irradiation and improving the safety and effectiveness of treatment. Through patterned irradiation using a digital micromirror device (DMD), the control system can quickly and accurately deliver UVA energy to various parts of the cornea, effectively shortening treatment time and improving treatment efficiency.

[0022] In some embodiments, a multimodal data fusion module is used to accurately fuse corneal multimodal data. The multimodal data fusion module includes a curvature processing unit, an OCT analysis unit, a biomechanical mapping unit, and a feature fusion unit. The curvature processing unit is configured to apply a Gaussian filter with a value of σ = 1.5 to the corneal topography curvature matrix to eliminate motion artifacts caused by eye movement or device noise, while preserving curvature mutation features (such as the steep gradient in the inferior temporal quadrant of keratoconus). The filter is then normalized to the interval [-1, 1] to generate a 512×512 pixel curvature tensor. Patient-adaptive curvature normalization avoids group statistical bias and enhances the sharpness of the features in the steep region. The OCT analysis unit is configured to automatically segment the corneal epithelium, stroma, and endothelium from the OCT volume data using a pretrained U-Net neural network. The unit then outputs the boundary coordinates of each layer and calculates a stromal thickness distribution map (in μm) based on the segmentation results. This map is then registered to the coordinate system of the curvature tensor using an affine transformation, ensuring spatial alignment within 50 μm (meeting the ISO 24294 standard for ophthalmic image alignment). The biomechanical mapping unit is configured to convert the biomechanical parameters obtained from the measurements into a 512×512 pixel biomechanical constraint matrix, maintaining the same spatial resolution as the curvature data. Examples of these biomechanical parameters include the first applanation deformation amplitude (SP-A1, in mm) and the corneal radius of curvature (HC radius, in mm) obtained from the Corvis ST measurement. The feature fusion unit is configured to concatenate the curvature tensor, matrix thickness distribution map, biomechanical constraint matrix, and lesion probability mask along the channel dimension to form a four-channel fusion tensor (i.e., a 512×512×4-dimensional fusion tensor), and perform Z-score normalization on each channel.

[0023] This application combines the curvature information of corneal topography, the tomographic structure information of OCT, biomechanical parameters and lesion probability mask to more comprehensively evaluate the state of the cornea and avoid the limitations of single modality data. The generation process of the four-channel fusion tensor is as follows: Figure 2 shown.

[0024] In some embodiments, the feature fusion unit also includes a risk area enhancer module for automatically identifying and enhancing high-risk corneal lesion areas, such as the inferior temporal quadrant of keratoconus. The risk area enhancer module is configured to: identify qualified pixel areas based on a preset normalized thickness threshold, normalized curvature threshold, and SP-A1 mapping threshold; and multiply the fusion tensor eigenvalue corresponding to the pixel area by an enhancement coefficient.

[0025] Specifically, for example, the preset normalized thickness threshold is -0.8. When the normalized thickness value is less than -0.8, the corresponding physical thickness is <400μm; the preset normalized curvature threshold is 0.6. When the normalized curvature value is greater than 0.6, it is significantly steep, which is a typical feature of keratoconus; the SP-A1 mapping threshold is 0.25. When the SP-A1 mapping value is greater than 0.25, the mechanical properties are weak and the risk of deformation is high.

[0026] The fused tensor eigenvalues ​​corresponding to pixel regions meeting the above conditions are multiplied by an enhancement factor of 1.2 to 1.8, giving them a higher weight in subsequent virtual modeling. This enhanced feature significantly influences the attention distribution of the cGAN generator and the gradient distribution of the energy heatmap, improving the generated resolution of high-risk areas and automatically increasing the UVA energy density in these areas by 15 to 30%.

[0027] In some embodiments, the dynamic energy planning module is configured to: apply an energy attenuation coefficient to areas meeting a stromal thickness threshold in the virtual corneal model; and generate a light spot arrangement pattern with a light spot diameter of 50-100 μm and a light spot density in the center area 1.5-2 times that in the edge area. The light spot can be, for example, a honeycomb or gradient ring pattern.

[0028] For regions meeting the stromal thickness threshold in the virtual corneal model, for example, those with a stromal thickness <450μm, the energy attenuation coefficient k is applied as follows: k = 0.9 - 1.3e^{-0.01d}, where d is the Euclidean distance from the current voxel to the thinnest point, expressed in μm. For example, at the thinnest point (d = 0), k ≈ 0.3 (energy reduced to 30% to avoid perforation); at d = 100μm, k ≈ 0.6 (energy restored to 60%); and at d ≥ 300μm, k approaches 0.9 (retaining a 10% attenuation margin). The automatic intraoperative dynamic energy planning module adjusts the attenuation coefficient in real time based on thickness changes detected by OCT.

[0029] Specifically, the virtual cornea model generation flow chart proposed in one embodiment of the present application can be referred to Figure 3 .

[0030] In some embodiments, the control system also includes a dynamic optimization module configured to monitor riboflavin fluorescence intensity in real time and update the illumination strategy when the local riboflavin fluorescence intensity falls below the global mean by more than a preset percentage. Optionally, an embedded spectral sensor can monitor riboflavin fluorescence intensity in real time and convert the light intensity signal into a two-dimensional distribution graph. The embedded spectral sensor has a detection wavelength of 365nm ± 5nm (riboflavin excitation peak), a signal-to-noise ratio > 50dB, a spatial resolution of 100μm, a sampling rate of 10Hz, and covers an illumination area of ​​6mm in diameter. For example, if the local riboflavin fluorescence intensity falls below the global mean by more than 30%, the local area is identified as a low-reactivity region, and the illumination strategy can be updated using a Q-learning algorithm. For low-reactivity regions, the illumination strategy can be updated by increasing the energy density (up to +30%), extending the illumination time (up to +15s), or switching the spot pattern (e.g., from honeycomb to spiral scanning).

[0031] In some embodiments, the control system also includes a motion compensation module, which solves the core problem of spot drift caused by micro-movement of the eye during surgery through a sub-millisecond dynamic tracking-compensation system, and improves error control from millimeter level to micron level. Figure 4 As shown, the motion compensation module is configured to: track pupil edge feature points and calculate the eye deflection vector; establish an eye-robotic arm motion model based on the eye deflection vector; drive the lifting track and robotic arm of the corneal cross-linking device to complete synchronous offset within a preset time based on the eye-robotic arm motion model; and make the projection position error of the light spot ≤50μm through closed-loop control.

[0032] Specifically, a PID control loop can be used for closed-loop control. In one embodiment, the motion compensation module includes a visual unit, such as a 200fps infrared camera (resolution 1280×1024), combined with an 850nm ring light source, to track 6-8 feature points on the pupil edge. The infrared camera acquires the pupil feature point coordinates every 5ms and calculates the eye deflection vector (accuracy ±0.02°) by ellipse fitting; according to Δx eye Deviation from the target position (initial calibration point) e(t) , real-time update of PID output signal u(t) .Will u(t) Converted to joint torque command (via Jacobian matrix J −1 The system uses a CNC mapping algorithm to control the X / Y / θ-axis motors to complete compensation movements within 50ms. The CNC lift rail (Z-axis) compensates for longitudinal displacement, while the rigid robotic arms (X / Y / θ) compensate for lateral offset and rotation to accommodate complex eye movements (such as Bell's phenomenon). The next frame of camera data verifies the residual error after compensation. If it exceeds 50μm, a secondary correction (adaptive PID) is triggered.

[0033] In some embodiments, the control system also includes a three-level safety fuse module, comprising a primary protection unit, a secondary protection unit, and a tertiary protection unit. The primary protection unit is configured to detect UVA light density fluctuations in real time and trigger automatic calibration when the fluctuation amplitude exceeds a fluctuation threshold (e.g., ±10%). For real-time detection of UVA light density fluctuations, for example, a UV light density sensor (wavelength 365nm) sampling at a 1kHz frequency can be used to detect fluctuations. Automatic calibration can utilize a closed-loop PID calibration, which can restore light intensity to a set value within 200ms to avoid energy shortage or overexposure. The secondary protection unit is configured to shut off the UVA light source and lock the robotic arm travel when the distance sensor detects that the distance between the corneal cross-linking device's probe and the cornea is less than a distance threshold (e.g., <5mm). For example, an infrared time-of-flight sensor (accuracy ±0.1mm) can be used to detect the probe-cornea distance. A collision risk is determined when the secondary protection unit detects a probe-cornea distance less than 5mm, at which point it shuts off the light source and locks the robotic arm travel. The arm's travel is locked by an electromagnetic brake, which locks the arm's travel within 10 milliseconds upon receiving a signal. The Level 3 protection unit is configured to trigger a digital micromirror device (DMD) to dynamically attenuate the abnormal area's UV light energy by 0.3-0.5 (reducing the 365nm UV light energy from the standard 5.4J / cm² to 1.6-2.7J / cm²) if the intraoperative OCT detects a localized reduction in stromal thickness greater than a preset percentage (e.g., 10%). Simultaneously, the balancing fluid perfusion system is activated to promote corneal recovery to a safe threshold. The OCT updates the full corneal thickness map every 30 seconds to detect localized dehydration or excessive thinning.

[0034] This three-level safety fuse module upgrades passive alarms to active protection, improving clinical safety by 90%. Through a hierarchical and progressive protection mechanism, it addresses the multi-dimensional safety risks during corneal cross-linking surgery.

[0035] In some embodiments, the control system further includes an intraoperative adjustment module, which uses dynamic energy regulation driven by real-time data to address the problem of planned-actual thickness deviation caused by corneal dehydration, uneven riboflavin penetration, etc. during surgery. Specifically, the intraoperative adjustment module is configured to obtain the actual thickness distribution T_real(t) detected by intraoperative OCT at a preset interval (e.g., 60 seconds). When the deviation between the actual thickness distribution T_real(t) and the thickness predicted by the virtual corneal model T_model exceeds a threshold, the applied energy is adjusted according to the following formula: ΔE=0.5×|T_{real}(t)-T_{model}| / T_{model}\quad (J / cm²) When ΔE>0.3J / cm², energy distribution recalculation is triggered.

[0036] The coefficient 0.5 in the formula is determined through finite element optimization to balance response speed and stability.

[0037] Through the setting of the above-mentioned intraoperative adjustment module, the energy can be automatically increased by 7.5% (ΔE=0.375 J / cm²) for thickness reduction caused by dehydration (such as 400μm→370μm); for abnormally thickened areas (such as edema), the energy can be intelligently reduced to avoid over-irradiation.

[0038] In some embodiments, the control system further includes a postoperative evaluation module, which is configured to: compare changes in corneal biomechanical index before and after surgery; and generate a secondary treatment recommendation report when the corneal biomechanical index decreases by less than an index threshold after a preset time after surgery. Specifically, taking the Corvis ST system as an example, the stress-strain relationship analysis of the Corvis ST is used to identify whether the postoperative corneal deformation resistance is maintained at a safe threshold (such as a critical value of a biomechanical index decrease of ≥30%). When the corneal biomechanical index decreases by <30% after a preset time after surgery (such as 6 months), a secondary treatment recommendation report is generated. The present application can provide early warning of corneal ectasia risks, trigger an intervention mechanism when biomechanical preservation is insufficient, and detect subclinical abnormalities 3-6 months earlier than traditional topography screening.

[0039] The second aspect of the present invention provides a method for controlling a corneal cross-linking device, which is implemented by the control system described in any of the above embodiments, such as Figure 5 As shown, the method includes the following steps: integrating the preoperative corneal topography curvature matrix, OCT layered thickness data and biomechanical parameters to generate a four-channel fusion tensor; using the four-channel fusion tensor as a condition, inputting a conditional generative adversarial network to generate a virtual corneal model; based on the mechanical response prediction of the virtual corneal model, generating a thermal map of energy distribution; and performing patterned irradiation based on the thermal map.

[0040] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0041] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A control system for a corneal cross-linking device, characterized in that: include: The multimodal data fusion module is configured to integrate the preoperative corneal topography curvature matrix, OCT layered thickness data, and biomechanical parameters to generate a four-channel fusion tensor; A virtual cornea modeling module is configured to generate a virtual cornea model by inputting a conditional generative adversarial network based on the four-channel fusion tensor; a dynamic energy planning module configured to generate a thermal map of energy distribution based on the mechanical response prediction of the virtual corneal model; The irradiation control module is configured to perform patterned irradiation based on the heat map.

2. The control system according to claim 1, characterized in that: The multimodal data fusion module includes: a curvature processing unit configured to perform Gaussian filtering and normalization on the preoperative corneal topography curvature matrix to generate a standardized curvature tensor; an OCT analysis unit configured to segment the corneal epithelium, stroma, and endothelium in the OCT volume data using a pre-trained neural network, calculate a stroma thickness distribution map, and register it to the coordinate system of the curvature tensor; a biomechanical mapping unit, which converts the biomechanical parameters obtained by the detection into a biomechanical constraint matrix, wherein the biomechanical parameters include at least the amplitude of the first applanation deformation and the curvature radius of the most concave point of the cornea; The feature fusion unit is configured to splice the curvature tensor, the matrix layer thickness distribution map and the biomechanical constraint matrix along the channel dimension to form a four-channel fusion tensor, and perform standardization processing on each channel.

3. The control system according to claim 2, characterized in that: The feature fusion unit further includes a risk region enhancement submodule configured to: Identify pixel areas that meet the conditions based on a preset normalized thickness threshold, a normalized curvature threshold, and a first flattening deformation amplitude mapping threshold; The fused tensor eigenvalue corresponding to the pixel area is multiplied by the enhancement coefficient.

4. The control system according to claim 2, characterized in that: The dynamic energy planning module is configured to: Applying an energy attenuation coefficient to an area meeting a stromal layer thickness threshold in the virtual corneal model; A light spot arrangement pattern is generated, wherein the diameter of the light spot is 50-100 μm, and the light spot density in the central area is 1.5-2 times that in the edge area.

5. The control system according to claim 1, characterized in that: Also included is a dynamic optimization module, configured to: Real-time monitoring of riboflavin fluorescence intensity; When the local riboflavin fluorescence intensity is lower than the average value of the entire field area by more than a preset ratio, the irradiation strategy is updated.

6. The control system according to claim 4, characterized in that: Also included is a motion compensation module configured to: Track pupil edge feature points and calculate eye deflection vector; Establishing an eyeball-robotic arm motion model according to the eyeball deflection vector; Based on the eyeball-robotic arm motion model, the lifting track and the robotic arm are driven to complete synchronous offset within a preset time; The projection position error of the light spot is made ≤50 μm through closed-loop control.

7. The control system according to claim 2, characterized in that: It also includes a three-level safety fuse module, which includes: The first-level protection unit is configured to detect UVA light density fluctuations in real time and trigger automatic calibration when the fluctuation amplitude exceeds the fluctuation threshold; The secondary protection unit is configured to cut off the UVA light source and lock the robotic arm travel when the distance sensor detects that the probe-cornea distance is less than a distance threshold; The third-level protection unit is configured to automatically start the energy attenuation mode of the local area when the intraoperative OCT detects that the local thickness of the matrix layer is reduced by more than a preset ratio, with an attenuation coefficient of 0.3~0.

5.

8. The control system according to claim 2, characterized in that: Also included is an intraoperative adjustment module configured to: The actual stromal thickness distribution detected by intraoperative OCT is obtained at preset time intervals. When the deviation between the actual stromal thickness and the predicted stromal thickness of the virtual corneal model exceeds a threshold, the applied energy is adjusted.

9. The control system according to claim 1, characterized in that: Also included is a post-operative assessment module configured to: Comparison of changes in corneal biomechanical index before and after surgery; When the corneal biomechanical index decreases below the index threshold after the preset time after surgery, a secondary treatment recommendation report is generated.

10. A method for controlling a corneal cross-linking device, characterized in that: The method is implemented by a control system according to any one of claims 1 to 9, comprising the following steps: Integrate preoperative corneal topography curvature matrix, OCT layer thickness data and biomechanical parameters to generate a four-channel fusion tensor; Taking the four-channel fusion tensor as a condition, inputting a conditional generative adversarial network to generate a virtual cornea model; generating a thermal map of energy distribution based on the mechanical response prediction of the virtual corneal model; Patterned irradiation is performed based on the heat map.

Images