Search

CN-122023411-A - OCT-based cornea shaping mirror production quality assessment method and device

CN122023411ACN 122023411 ACN122023411 ACN 122023411ACN-122023411-A

Abstract

The invention discloses a cornea shaping lens production quality evaluation method and a device based on OCT, relating to the cornea shaping lens quality evaluation field, the method firstly obtains high-precision tomographic image data of the cornea shaping lens, and extracting pixel-level profiles of the tomographic images, performing sagittal calculation and three-dimensional reconstruction based on chord length definition, and finally constructing a visual evaluation system and evaluating production quality. The invention can improve the accuracy and efficiency of the production quality evaluation without collecting the fitting information of each lens of the wearer, and is further beneficial to improving the quality evaluation efficiency of the production of the same-series cornea shaping lens, thereby being beneficial to the popularization and application of the cornea shaping lens production standardization and cornea shaping lens production quality evaluation technology.

Inventors

  • JIA GANGYONG
  • LI MENGQI
  • XIE HONGBIAO
  • Ye Moujian
  • LIU YI
  • XU PEIFANG
  • YE JUAN
  • AN LIN

Assignees

  • 杭州电子科技大学

Dates

Publication Date
20260512
Application Date
20260413

Claims (16)

  1. 1. The cornea shaping mirror production quality assessment method based on OCT is characterized by comprising the following steps of: The method comprises the steps of 1, obtaining high-precision tomographic image data of a cornea shaping lens, specifically, placing the cornea shaping lens to be measured on a bearing platform, aligning by utilizing a sweep-source optical coherence tomography system or a spectral domain optical coherence tomography system and matching with a high-precision displacement table, executing a radial rotation scanning strategy, obtaining an original interference signal covering the whole diameter of the cornea shaping lens, and reconstructing the original interference signal into a two-dimensional tomographic image set through photoelectric conversion and signal processing; Step 2, carrying out pixel-level contour extraction on the tomographic image, namely constructing a contour recognition model based on a dual-stage intelligent segmentation architecture, generating and learning supervised features by combining an unsupervised mask, segmenting a lens region, analyzing the edges of the segmentation region based on differential geometric features, locking physical edge points, introducing a morphological constraint mechanism based on a thickness physical threshold, correcting artifact interference, and extracting front and rear surface contour point cloud data of the lens with topological continuity; Performing sagittal calculation and three-dimensional reconstruction based on chord length definition, specifically, performing central artifact restoration on the contour point cloud data, establishing a geometric vertex reference, performing spatial posture correction to eliminate acquisition inclination and scanning distortion, performing chromatographic slicing on the corrected three-dimensional model along the optical axis direction, calculating chord length values at different chromatographic depths, and establishing chord length-sagittal corresponding data streams within the full diameter range; The step 4 is to construct a visual evaluation system and evaluate the production quality, and the method comprises the steps of mapping the chord length-sagittal height corresponding data flow into a two-dimensional curve map, identifying structural defects by analyzing topological intersection relation between an actual measurement curve and a standard curve, and quantitatively evaluating the resolution and the stability of the production process by utilizing a differential variance model of lenses with adjacent specifications.
  2. 2. The OCT-based cornea shaping mirror production quality assessment method is characterized in that an annular supporting structure is adopted as an objective table by the bearing platform, the inner diameter of the annular supporting structure is smaller than the overall diameter of the cornea shaping mirror to be measured, a non-optical area at the edge of a lens is used for physical supporting, suspension and minimum contact fixation of an optical area of the cornea shaping mirror are achieved, the central area of the bearing platform is designed to be a hollow structure, a plurality of layers of fine ventilation holes are arranged on the side wall or the bottom of the annular supporting structure, and black anodized aluminum or light-absorbing polymers subjected to matte treatment are selected as the bottom and contact surface materials of the bearing platform.
  3. 3. The method for evaluating the production quality of the cornea shaping mirror based on OCT according to claim 1 is characterized in that the high-precision displacement table integrates three-dimensional fine adjustment of an X axis, a Y axis and a Z axis and a plane tilt function, wherein the X axis and the Y axis are used for adjusting the geometric center of the cornea shaping mirror to be precisely aligned with the optical axis of an OCT system, the Z axis is used for adjusting and matching with the focal length adjustment of an optical system, and the plane tilt function allows a bearing platform to carry out fine adjustment on the sample posture in an orthogonal direction.
  4. 4. The method for evaluating the production quality of the OCT-based cornea shaping mirror is characterized in that in the step 1, the radial rotation scanning strategy is implemented by controlling a scanning probe of OCT equipment to take the geometric center of a lens as a scanning origin, driving a two-dimensional scanning galvanometer to enable scanning beams to conduct linear scanning along a preset radial path, setting the number of the scanning lines to be N, uniformly distributing the N scanning lines on the circumference of 360 degrees at equal angle intervals, and enabling all the scanning lines to meet at the scanning origin to form high-density sampling coverage of the central area of the lens.
  5. 5. The method for evaluating the production quality of the cornea shaping lens based on OCT according to claim 1, wherein in the step 2, the contour recognition model based on a dual-stage intelligent segmentation architecture is constructed into two stages, wherein the first stage is unsupervised mask generation, wherein a clustering algorithm based on region growth or an adaptive statistical threshold method is utilized to analyze gray histogram distribution and local texture difference of an OCT image, a lens region with high signal intensity and a background region with low signal intensity are distinguished, a binary mask comprising a lens main body form is generated, the second stage is supervised feature learning, the binary mask is used as a pseudo tag or a spatial attention guide graph, an input model is trained, and the model is configured to learn high-dimensional semantic features for distinguishing lens boundary signals and background noise.
  6. 6. The method for evaluating the production quality of the cornea shaping lens based on OCT according to claim 5 is characterized in that in the step 2, the edges of the divided areas are analyzed based on differential geometric features, the locked physical edge points are specifically obtained by carrying out path tracking on potential profile curves divided by the model, calculating the first derivative and curvature of the tangential slope along the profile path in real time, detecting the abrupt change points of the first derivative of the slope or extreme points in a curvature map, judging the abrupt change points or the extreme points as the left and right actual physical edge points of the lens by utilizing the cutoff or the rounded geometric features existing on the edges of the cornea shaping lens, cutting out the effective lens physical areas according to the physical edge points, and eliminating invalid background noise signals on two sides.
  7. 7. The method for evaluating the production quality of the cornea shaping mirror based on OCT is characterized in that in the step 3, center artifact restoration is carried out on the contour point cloud data, the geometric vertex standard is specifically established by identifying a center highlight saturated area or a data missing area caused by vertical incidence light in an OCT image, setting a mask radius with a scanning center as an origin, eliminating original low-confidence point cloud data in the radius range to form a center cavity, extracting high-quality contour data in an annular area close to the edge of the center cavity, selecting a fitting model according to the design type of the cornea shaping mirror, carrying out regression fitting on a spherical mirror design by adopting a spherical equation, carrying out regression fitting on a diffuse mirror design by adopting an ellipsoidal equation, calculating theoretical height values of coordinate points in the cavity area by utilizing a mathematical model obtained by fitting, and backfilling the reconstructed points into three-dimensional data sets, so that the precise geometric vertex standard for vector height calculation is established.
  8. 8. The method for evaluating the production quality of the cornea shaping mirror based on OCT according to claim 7 is characterized in that in the step 3, the spatial posture correction comprises a rapid reconstruction path based on two-dimensional plane rotation, specifically, for each tomographic image, two physical boundary points are identified, two boundary points are connected, a midpoint of a connecting line is taken as a local geometric center projection of the cross section, a straight line which passes through the midpoint and is perpendicular to the connecting line of the boundary points is taken as a local central axis, an included angle between the local central axis and a vertical axis of an image pixel coordinate system is calculated, the image center is taken as a rotation base point, a two-dimensional rotation matrix is applied to rotate the cross section image until the connecting line of the two boundary points is horizontal and the central axis is vertical, and all rotated outlines are spliced under a cylindrical coordinate system according to an acquisition angle by taking the central axis of each cross section as a unified Z-axis reference.
  9. 9. The method for evaluating the production quality of the cornea shaping mirror based on OCT (optical coherence tomography) is characterized by comprising the steps of performing spatial attitude correction in step 3, including a global reconstruction path based on three-dimensional spatial rotation, specifically, applying a pre-calibrated pixel offset compensation value to each scanning line before contour stitching aiming at radial dislocation caused by mechanical inertia or electronic delay of a scanning vibrating mirror through radial scanning fixed value offset correction, performing surface fitting, extracting boundary point sets of all scanning contour ends, performing surface fitting on the boundary point sets by adopting a least square method, performing iterative fitting for the second time after removing outliers if residual errors exceed a threshold value to obtain a lens bottom equation, calculating a global rotation pair, calculating a normal vector of a fitting bottom surface, constructing a rotation matrix, rotating the normal vector to be parallel to a Z axis of a measurement coordinate system, and accordingly calibrating the lens bottom surface to an XOY plane to realize the alignment of a geometric optical axis and the Z axis.
  10. 10. The method for evaluating the production quality of an OCT-based cornea shaping lens according to claim 1, wherein in the step 3, the tomographic slice is performed on the corrected three-dimensional model along the optical axis direction, and the chord length values at different tomographic depths are calculated by establishing a coordinate system with a lens vertex as an origin and the corrected optical axis as a Z-axis, starting from the lens vertex, and taking a micrometer-scale step length along the negative Z-axis direction The method comprises the steps of discretizing slices, intercepting a cross section profile point cloud of the rear surface of a lens at the height of each slice, calculating the average Euclidean distance from the cross section profile point cloud to the center of an optical axis for a cornea shaping mirror of conventional rotationally symmetrical design, defining the average distance as a chord length radius corresponding to the sagittal height, fitting the cross section profile point cloud for the cornea shaping mirror of astigmatic design by adopting a least square ellipse fitting algorithm, extracting a long half axis and a short half axis of a fitting ellipse, defining the long half axis as the chord length of a flat meridian direction at the sagittal height, and defining the short half axis as the chord length of a steep meridian direction at the sagittal height.
  11. 11. The method for evaluating the production quality of the cornea shaping mirror based on OCT according to claim 1 is characterized in that in the step 4, the chord length-sagittal height corresponding data flow is mapped into a two-dimensional curve map, specifically, a rectangular coordinate system is established by taking a chord length value as an abscissa and a corresponding sagittal height value as an ordinate, and discrete chord length-sagittal height data points are subjected to spline interpolation processing for three times to generate a smooth curve with continuous first-order and second-order derivatives, and the smooth curve intuitively shows the depth distribution rule of the cornea shaping mirror in the whole diameter range from the geometric center to the edge, wherein the depth distribution rule comprises the flattening amount of an optical zone, the cutting depth of a reversal zone and the tilting height of a peripheral zone.
  12. 12. The method for evaluating the production quality of the OCT-based cornea shaping lens according to claim 1 is characterized in that in the step 4, the structural defect is identified by analyzing the topological intersection relation between an actually measured curve and a standard curve, specifically, the method comprises the steps of obtaining standard design curves or statistical envelope curves of the same batch or series, superposing an actually measured chord length-sagittal height curve of a lens to be tested and the standard curve under the same coordinate system, detecting whether unexpected intersection points exist between the two curves, and judging that the lens has anisotropic processing errors if the intersection points exist.
  13. 13. The method for evaluating the production quality of the cornea shaping lens based on OCT is characterized in that in the step 4, a differential variance model of lenses with adjacent specifications is utilized, the resolution and stability of a production process are evaluated quantitatively, specifically, a reference lens with the lens to be tested being close to design parameters in the same series is selected, the reference lens is a lens with a base arc adjacent first grade or a base arc adjacent first grade, a chord length-sagittal height data sequence corresponding to the reference lens is taken in a full diameter range, a sagittal height difference value set of the reference lens and the reference lens at all aligned chord length positions is calculated, a statistical variance or standard deviation of the difference value set is calculated, the variance is used as an index for measuring the stability of the processing difference, if the variance value is lower than a preset stability threshold, the production line is judged to have effective process resolution, the processing difference is uniform, and if the variance value is higher than the preset threshold, the production line is judged to have random errors or insufficient resolution, and adjacent tiny indexes cannot be effectively distinguished physically.
  14. 14. The method for evaluating the production quality of the cornea shaping lens based on OCT (optical coherence tomography) according to claim 1 is characterized by further comprising the steps of constructing a multi-dimensional index-oriented batch-level quality grading system, specifically, converging chord length-sagittal height curves of all lenses in the same production batch, constructing a statistical envelope interval, calculating the structural dispersion of the batch, combining curve form deviation, topological intersection condition and adjacent specification difference variance indexes to generate a comprehensive quality score, and mapping the comprehensive quality score into specific technical problems including mold abrasion degree, peripheral cutting unbalance or center fixed point offset.
  15. 15. OCT-based cornea shaping lens production quality assessment apparatus for performing the OCT-based cornea shaping lens production quality assessment method according to any one of claims 1 to 14, characterized in that the apparatus comprises: the acquisition module is used for acquiring an OCT image of the cornea shaping lens to be tested; The calculation module is used for determining target lens parameters of the cornea shaping lens according to the OCT image of the cornea shaping lens, wherein the target lens parameters of the cornea shaping lens comprise chord length parameters corresponding to each sagittal parameter of the cornea shaping lens, all the sagittal parameters are used for indicating the vertical distance from the central top point of the inner surface of the cornea shaping lens to any reference horizontal plane of the cornea shaping lens, and the chord length corresponding to each sagittal parameter is used for indicating the distance from the edge point of the inner surface of the cornea shaping lens passing through the central point of the reference horizontal plane in the reference horizontal plane corresponding to the sagittal parameter; and the evaluation module is used for performing cornea shaping lens generation quality evaluation operation on the cornea shaping lens of the same series based on various data of the cornea shaping lens.
  16. 16. An OCT-based corneal shaping lens production quality assessment system, the system comprising: a memory storing executable program code; a processor coupled to the memory; the processor invokes the executable program code stored in the memory to perform an OCT-based cornea shaping lens production quality assessment method of any one of claims 1 to 14.

Description

OCT-based cornea shaping mirror production quality assessment method and device Technical Field The invention relates to the field of cornea shaping mirror quality evaluation, in particular to a cornea shaping mirror production quality evaluation method and device based on OCT. Background In the field of ophthalmic medical treatment, a cornea shaping lens has been widely applied, temporary reshaping of cornea morphology is realized through mechanical compression and tear fluid pressure action during night wearing, and the effect of clear naked eye vision can be achieved without wearing glasses in the daytime, wherein the cornea shaping lens is used as a third type of medical instrument directly contacting cornea, the quality of the cornea shaping lens is directly related to eye safety, and the cornea shaping lens production quality assessment method is particularly important in order to avoid risks of cornea damage, infection and the like caused by material defects and parameter deviation. The traditional evaluation method mainly starts from three aspects of product materials, product design and processing technology, the production precision and quality internal control of manufacturers are not paid attention to, the traditional production quality evaluation method mostly needs to collect and analyze a large amount of eye data before and after wearing by a cornea shaping lens wearer, and the wearer usually needs to try on various different types of lenses, so that the cornea of the wearer is easily damaged while a large amount of manpower and time cost are consumed by the evaluation mode, and the standardized process of cornea shaping lens production evaluation is limited. It is important to provide a method for improving the accuracy and efficiency of the quality assessment of the cornea shaping lens production. Disclosure of Invention The invention provides a cornea shaping lens production quality assessment method based on OCT and chord length sagittal height calculation, which can improve the accuracy and efficiency of production quality assessment without collecting the trial wearing information of each lens of a wearer, and is further beneficial to improving the quality assessment efficiency of the same-series cornea shaping lens production, thereby being beneficial to the popularization and application of cornea shaping lens production standardization and cornea shaping lens production quality assessment technology. In order to achieve the above object, the present application provides the following solutions: In a first aspect, the present invention provides a method for evaluating production quality of an OCT-based cornea shaping lens, comprising the steps of: The method comprises the steps of 1, obtaining high-precision tomographic image data of a cornea shaping lens, specifically, placing the cornea shaping lens to be measured on a bearing platform, aligning by utilizing a sweep-source optical coherence tomography system or a spectral domain optical coherence tomography system and matching with a high-precision displacement table, executing a radial rotation scanning strategy, obtaining an original interference signal covering the whole diameter of the cornea shaping lens, and reconstructing the original interference signal into a two-dimensional tomographic image set through photoelectric conversion and signal processing; Step 2, carrying out pixel-level contour extraction on the tomographic image, namely constructing a contour recognition model based on a dual-stage intelligent segmentation architecture, generating and learning supervised features by combining an unsupervised mask, segmenting a lens region, analyzing the edges of the segmentation region based on differential geometric features, locking physical edge points, introducing a morphological constraint mechanism based on a thickness physical threshold, correcting artifact interference, and extracting front and rear surface contour point cloud data of the lens with topological continuity; Performing sagittal calculation and three-dimensional reconstruction based on chord length definition, specifically, performing central artifact restoration on the contour point cloud data, establishing a geometric vertex reference, performing spatial posture correction to eliminate acquisition inclination and scanning distortion, performing chromatographic slicing on the corrected three-dimensional model along the optical axis direction, calculating chord length values at different chromatographic depths, and establishing chord length-sagittal corresponding data streams within the full diameter range; The step 4 is to construct a visual evaluation system and evaluate the production quality, and the method comprises the steps of mapping the chord length-sagittal height corresponding data flow into a two-dimensional curve map, identifying structural defects by analyzing topological intersection relation between an actual measurement curve and a standa