Eyeball calibration method based on visual image processing technology

Through visual image processing technology, real-time capture and analyze eye dynamic images, the problem of low eye tracking accuracy is solved, and diagnostic assistance is provided for the health status of the eye, helping medical staff to quickly and accurately evaluate eye diseases.

CN119090835APending Publication Date: 2024-12-06BOCK MEDICAL TECH (SHANGHAI) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411167756.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing eye tracking model has low accuracy in the tracking results of eye position when the head is shaking relative to the acquisition device, and lacks technology based on eye dynamic analysis to assist medical staff in assessing the patient's eye disease recovery status.

Method used

Using eyeball calibration method based on visual image processing technology, the eyeball dynamic images are captured in real time through high-speed cameras, the eyeball picture frame is decomposed and marked, the tracking points and motion paths are set, the difference in motion paths is analyzed, and the eyeball health status is judged based on the health situation judgment logic. .

Benefits of technology

It provides an effective auxiliary reference for the user's eye health status, helping medical staff to conduct rapid and accurate diagnosis of eye diseases.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119090835A_ABST
    Figure CN119090835A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data image processing, in particular to an eyeball calibration method based on a visual image processing technology, which comprises the following steps: capturing a user eyeball dynamic image in real time through a high-speed camera, and decomposing the user eyeball dynamic image to obtain a user eyeball picture frame; performing corresponding source timestamp marking and storage on the user eyeball picture frames obtained through decomposition; the stored user eyeball picture frames are selected, the user eyeball picture frames are further decomposed in the user eyeball dynamic image through collection of the user eyeball dynamic image, the user eyeball movement path is obtained based on tracking point setting, and finally the user eyeball movement path is obtained based on continuous obtaining and difference analysis of the user eyeball movement path. According to the technical scheme, effective auxiliary reference is provided for judging the health state of the eyeballs of the user, so that more effective auxiliary reference data can be provided when medical staff diagnose the eye problems of the user, and the eye problems of the user can be quickly and accurately diagnosed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data image processing, and in particular to an eyeball calibration method based on visual image processing technology. Background Art

[0002] Some diseases can cause obstruction of eye movement, such as thyroid-related eye disease. Abnormal thyroid function triggers an immune response, causing edema and fibrosis of the extraocular muscles. Orbital tumors can compress the extraocular muscles and nerves, causing movement disorders. When vascular lesions in the brain involve the oculomotor nerves, abnormal eye movement can also occur, affecting the normal function of the eyes.

[0003] The invention patent with application number 201810988937.2 discloses an eyeball position tracking method, which is characterized by comprising: processing the facial image of the person to be tested at the current moment, determining the rotation information of the eye image and the head in the facial image relative to the visual reference point, wherein the visual reference point is the initial position of the head in the facial image corresponding to the initial moment relative to the image acquisition device; inputting the rotation information of the eye image and the head relative to the visual reference point into the preset eye tracking type to determine the position of the eye relative to the visual reference point. The processing of the facial image of the person to be tested at the current moment, determining the rotation information of the eye image and the head in the facial image relative to the visual reference point, specifically comprises: dividing the facial image of the person to be tested at the current moment into feature areas to determine the eye image corresponding to the eye area; based on the predetermined standard coordinates of the head relative to the visual reference point, determining the offset coordinates of the rotation of the head in the facial image relative to the visual reference point as the rotation information of the head relative to the visual reference point.

[0004] This application aims to solve the problem that "the existing eye tracking model has low accuracy in tracking the eye position when the head shakes relative to the acquisition device."

[0005] However, for eye diseases, there is currently no technology based on ocular dynamic analysis to assist medical staff in assessing the recovery status of patients with eye diseases;

[0006] To this end, we propose an eye calibration method based on visual image processing technology. Summary of the invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides an eye calibration method based on visual image processing technology, which solves the technical problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] The eye calibration method based on visual image processing technology includes:

[0010] The user's eye dynamic images are captured in real time by a high-speed camera, the user's eye dynamic images are decomposed to obtain user eye picture frames, and the corresponding source timestamps are marked and stored for each decomposed user eye picture frame; the stored user eye picture frame is selected, the user's eye dynamic tracking area is intercepted in the selected user eye picture frame, and the tracking point is further set in the intercepted user eye dynamic tracking area; based on the connection of the tracking points in the two selected groups of user eye picture frames, the user's eye sub-motion paths are obtained in several groups, and the several groups of user eye sub-motion paths are connected to each other to obtain the user's eye movement path; the above steps are repeated to continuously obtain the user's eye movement path, and the difference of the continuously obtained user's eye movement path is analyzed; the above steps are repeated to continuously obtain the user's eye movement path, and the difference of the continuously obtained user's eye movement path is analyzed; the user's eye health status judgment logic is set, and whether the user's eye is healthy is determined according to the difference of the user's eye movement path combined with the user's eye health status judgment logic.

[0011] Furthermore, when the user eyeball dynamic image captured by the high-speed camera is decomposed into user eyeball picture frames, the number of the decomposed user eyeball picture frames is twice the duration of the user eyeball dynamic image, and each group of decomposed user eyeball picture frames is synchronously identified in the storage stage after marking the corresponding source timestamp in the user eyeball picture frame. The user eyeball area image identified in the user eyeball picture frame is segmented in the user eyeball picture frame, and the segmented user eyeball area image is marked with the source timestamp, and then further stored.

[0012] When determining the number of frames of the user's eyeball image, the unit of the duration of the user's eyeball dynamic image is defined as seconds.

[0013] Furthermore, the recognition logic of the user's eye area image is expressed as:

[0014] Two sets of user eye area image determination thresholds are set for the brightness and grayscale value determination of the user eye area, and the brightness and grayscale value of each pixel in the user eye image frame are further calculated:

[0015] ;

[0016] Where: is the pixel brightness; is the pixel gray value; , , The three sets of values ​​for the pixel based on the three color channels;

[0017] Among them, based on the above formula, the brightness and grayscale values ​​of each pixel in the user eyeball picture frame are calculated, and the calculation result is further compared with the corresponding user eyeball area image judgment threshold. The pixels whose brightness and grayscale values ​​meet the user eyeball area image judgment threshold are taken as non-segmentation targets, and the pixels whose brightness and grayscale values ​​do not meet the user eyeball area image judgment threshold are taken as segmentation targets. After the pixels in all user eyeball picture frames are judged as non-segmentation targets and segmentation targets, the user eyeball area image is segmented globally based on the user eyeball picture frame.

[0018] Furthermore, the operation of selecting the stored user eyeball picture frame and capturing the user eyeball dynamic tracking area in the selected user eyeball picture frame is, that is, selecting the user eyeball area image and capturing the user eyeball dynamic tracking area in the user eyeball area image;

[0019] The interception logic of the user's eye dynamic tracking area in the user's eye area image is expressed as:

[0020] A set of grayscale determination thresholds for the non-sclera area of ​​the user's eyeball is set, and the grayscale determination thresholds for the non-sclera area of ​​the user's eyeball are compared with pixels in the user's eyeball area image to obtain the non-sclera area of ​​the user's eyeball in the user's eyeball area image, and further obtain the center point of the non-sclera area of ​​the user's eyeball;

[0021] Analyze and select the similarity of the user's eye area images, and determine the user's eye dynamic tracking area based on the similarity and the center point of the user's eye non-sclera area;

[0022] Among them, the user's eye dynamic tracking area is a circular area with the center point of the user's eye non-sclera area as the center, and the diameter of the circular area is the product of the maximum inscribed circle diameter of the user's eye area image and the similarity of the user's eye area image.

[0023] Furthermore, the similarity analysis logic of the user's eye area image is expressed as:

[0024] ;

[0025] Where: is the similarity between the user's eyeball area image a and the user's eyeball area image b; , is the texture feature vector of the user's eyeball area image a and the user's eyeball area image b; , is the modulus of the two sets of vectors; , is the element value of the user eye area image a and the user eye area image b at the coordinate (i, j), where i and j represent the index of the row and column; , is the element average of the user's eye area image a and the user's eye area image b; , is the weight;

[0026] in, The value is between 0 and 1. The larger the value, the more similar the eye area images of the two groups of users are. Conversely, the smaller the value, the less similar the eye area images of the two groups of users are. , The sum of is 1, and < .

[0027] Furthermore, when performing the analysis operation on the similarity of the user's eye area images, two groups of user's eye area images are applied, and the two groups of user's eye area images are two groups of user's eye area images that are adjacent based on the marking timestamps;

[0028] The tracking point set in the user eye dynamic tracking area is a group of pixels with the smallest grayscale value in the user eye dynamic tracking area;

[0029] When the pixel with the minimum gray value is not unique, a group of pixels with the minimum gray value closest to the center of the user's eye dynamic tracking area is set as the tracking point.

[0030] Furthermore, based on the operation of connecting tracking points in the two selected groups of user eye area frames, that is, based on the operation of connecting tracking points in the two selected groups of user eye area images, the operation of selecting two groups of user eye area images to set tracking points and connecting tracking points to obtain the user eye movement path is performed continuously, and the group of user eye area images with the earliest marked timestamp in the next two groups of user eye area images selected and the group of user eye area images with the latest marked timestamp in the last two groups of user eye area images selected are the same group of user eye area images.

[0031] Furthermore, when the plurality of groups of user eyeball sub-motion paths are connected to each other, the timestamps of the user eyeball region images of the endpoints at both ends of each group of user eyeball sub-motion paths are synchronously identified, and the endpoints of the user eyeball sub-motion paths with the same timestamps are connected to each other to form a user eyeball movement path;

[0032] During the process of acquiring the user's eye movement path, the user's eye completes the eye movement based on a specified voice prompt, or completes the eye movement based on a specified dynamic target movement trajectory in an animation.

[0033] Furthermore, after the user's eye movement path is obtained, nodes on the user's eye movement path are further picked up, and each node is mapped to AutoCAD based on a 1:1 distance ratio between each other, and each group of mapping points is sequentially connected using a spline curve command to obtain a group of curves, and the obtained curves are used to iterate the originally obtained user's eye movement path.

[0034] Furthermore, when analyzing the differences in the eye movement paths of the users, the three groups of eye movement paths of the users that have been acquired the most recently are used for analysis. The three groups of eye movement paths of the users are recorded as x1, x2, and x3. The analysis logic is expressed as follows:

[0035] The user's eye health status determination logic is expressed as:

[0036] ;

[0037] Where: The differences in users’ eye movement paths; , is the number of control points on the eye movement paths of the two groups of users; is the u-th control point on the user's eye movement path x1; is the vth control point on the user's eye movement path x2; , The difference of the eye movement path of the first user and the difference of the eye movement path of the second user;

[0038] in, represents the magnitude of the distance between two points, Based on x1 and x2, Based on x2 and x3, formula (1) represents the logic of analyzing the difference of the user's eye movement path, and formula (2) represents the logic of judging the user's eye health status. When formula (2) holds, it means that the user's eye health status tends to recover or deteriorate. When formula (2) does not hold, it means that the current judgment result of the user's eye health status is invalid.

[0039] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0040] The present invention provides an eye calibration method based on visual image processing technology. During the execution of the method, the user's eye dynamic image is collected, and the user's eye picture frame is further decomposed in the user's eye dynamic image, so as to obtain the user's eye movement path based on the tracking point setting, and finally based on the continuous acquisition and difference analysis of the user's eye movement path, an effective auxiliary reference is provided for determining the health status of the user's eye, so that medical staff have more effective auxiliary reference data when diagnosing the user's eye problems, so as to make a quick and accurate diagnosis of the user's eye problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 Schematic diagram of the flow of the eye calibration method based on visual image processing technology. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] The present invention will be further described below in conjunction with the embodiments.

[0045] Embodiment 1:

[0046] The eye calibration method based on visual image processing technology in this embodiment is as follows: Figure 1 As shown, including:

[0047] Step 1: Capture the user's eyeball dynamic image in real time through a high-speed camera, decompose the user's eyeball dynamic image to obtain the user's eyeball picture frame, and mark and store the corresponding source time stamp for each decomposed user eyeball picture frame;

[0048] Step 2: Select a stored user eyeball picture frame, capture a user eyeball dynamic tracking area in the selected user eyeball picture frame, and further set a tracking point in the captured user eyeball dynamic tracking area;

[0049] Step 3: Based on the connection lines of the tracking points in the two selected groups of user eyeball picture frames, a plurality of groups of user eyeball sub-motion paths are obtained, and the plurality of groups of user eyeball sub-motion paths are connected to each other to obtain the user eyeball movement path;

[0050] Step 4: Repeat the above steps to continuously obtain the user's eye movement paths, and analyze the differences in the continuously obtained user's eye movement paths;

[0051] Step 5: Repeat the above steps to continuously obtain the user's eye movement paths, and analyze the differences in the continuously obtained user's eye movement paths;

[0052] Step 6: Set the user's eye health status judgment logic, and judge whether the user's eye is healthy based on the difference in the user's eye movement path and the user's eye health status judgment logic;

[0053] An operation of selecting a stored user eyeball picture frame and capturing a user eyeball dynamic tracking area in the selected user eyeball picture frame, that is, an operation of selecting a user eyeball area image and capturing a user eyeball dynamic tracking area in the user eyeball area image;

[0054] The interception logic of the user's eye dynamic tracking area in the user's eye area image is expressed as:

[0055] A set of grayscale determination thresholds for the non-sclera area of ​​the user's eyeball is set, and the grayscale determination thresholds for the non-sclera area of ​​the user's eyeball are compared with pixels in the user's eyeball area image to obtain the non-sclera area of ​​the user's eyeball in the user's eyeball area image, and further obtain the center point of the non-sclera area of ​​the user's eyeball;

[0056] Analyze and select the similarity of the user's eye area images, and determine the user's eye dynamic tracking area based on the similarity and the center point of the user's eye non-sclera area;

[0057] The user's eyeball dynamic tracking area is a circular area with the center point of the user's eyeball non-sclera area as the center, and the diameter of the circular area is the product of the maximum inscribed circle diameter of the user's eyeball area image and the similarity of the user's eyeball area image;

[0058] The similarity analysis logic of the user's eye area image is expressed as:

[0059] ;

[0060] Where: is the similarity between the user's eyeball area image a and the user's eyeball area image b; , is the texture feature vector of the user's eyeball area image a and the user's eyeball area image b; , is the modulus of the two sets of vectors; , is the element value of the user eye area image a and the user eye area image b at the coordinate (i, j), where i and j represent the index of the row and column; , is the element average of the user's eye area image a and the user's eye area image b; , is the weight;

[0061] in, The value is between 0 and 1. The larger the value, the more similar the eye area images of the two groups of users are. Conversely, the smaller the value, the less similar the eye area images of the two groups of users are. , The sum of is 1, and < ;

[0062] When analyzing the differences in user eye movement paths, the three groups of user eye movement paths obtained recently are used for analysis. The three groups of user eye movement paths are recorded as x1, x2, and x3. The analysis logic is expressed as follows:

[0063] The logic for determining the user's eye health status is expressed as:

[0064] ;

[0065] Where: The differences in users’ eye movement paths; , is the number of control points on the eye movement paths of the two groups of users; is the u-th control point on the user's eye movement path x1; is the vth control point on the user's eye movement path x2; , The difference of the eye movement path of the first user and the difference of the eye movement path of the second user;

[0066] in, represents the magnitude of the distance between two points, Based on x1 and x2, Based on x2 and x3, formula (1) represents the logic of analyzing the difference of the user's eye movement path, and formula (2) represents the logic of judging the user's eye health status. When formula (2) holds, it means that the user's eye health status tends to recover or deteriorate. When formula (2) does not hold, it means that the current judgment result of the user's eye health status is invalid.

[0067] In this embodiment, by executing the steps in the above method, a real-time calibration and tracking service is provided for the user's eye movements, which can be used as a reference for the user's eye health analysis, thereby bringing further health management effects to the user's eyes.

[0068] Embodiment 2:

[0069] In terms of specific implementation, based on Example 1, this example refers to Figure 1 The eyeball calibration method based on visual image processing technology in Example 1 is further described in detail:

[0070] When the user eyeball dynamic image captured by the high-speed camera is decomposed into user eyeball picture frames, the number of the decomposed user eyeball picture frames is twice the duration of the user eyeball dynamic image. In the storage stage after marking the corresponding source timestamp, each group of decomposed user eyeball picture frames synchronously identifies the user eyeball area image in the user eyeball picture frame, and the user eyeball area image identified in the user eyeball picture frame is segmented in the user eyeball picture frame. The segmented user eyeball area image is marked with the source timestamp and then further stored.

[0071] When determining the number of frames of the user's eyeball image, the unit of defining the duration of the user's eyeball dynamic image is seconds;

[0072] The recognition logic of the user's eye area image is expressed as:

[0073] Two sets of user eye area image determination thresholds are set for the brightness and grayscale value determination of the user eye area, and the brightness and grayscale value of each pixel in the user eye image frame are further calculated:

[0074] ;

[0075] Where: is the pixel brightness; is the pixel gray value; , , The three sets of values ​​for the pixel based on the three color channels;

[0076] Among them, based on the above formula, the brightness and grayscale values ​​of each pixel in the user eyeball picture frame are calculated, and the calculation result is further compared with the corresponding user eyeball area image judgment threshold. The pixels whose brightness and grayscale values ​​meet the user eyeball area image judgment threshold are taken as non-segmentation targets, and the pixels whose brightness and grayscale values ​​do not meet the user eyeball area image judgment threshold are taken as segmentation targets. After the pixels in all user eyeball picture frames are judged as non-segmentation targets and segmentation targets, the user eyeball area image is segmented globally based on the user eyeball picture frame.

[0077] In this embodiment, through the above settings, further execution logic support is provided for the execution of the steps of the method in Example 1, and the recognition logic of the specified user eye area image is provided.

[0078] Embodiment 3:

[0079] In terms of specific implementation, based on Example 1, this example refers to Figure 1 The eyeball calibration method based on visual image processing technology in Example 1 is further described in detail:

[0080] When performing the analysis operation on the similarity of the user's eyeball region images, two groups of user's eyeball region images are applied, and the two groups of user's eyeball region images are two groups of user's eyeball region images that are adjacent based on the marking timestamps;

[0081] The tracking point set in the user's eye dynamic tracking area is a group of pixels with the smallest gray value in the user's eye dynamic tracking area;

[0082] When the pixel with the minimum gray value is not unique, a group of pixels with the minimum gray value closest to the center of the user's eye dynamic tracking area is set as the tracking point;

[0083] Based on the operation of connecting the tracking points in the two selected sets of user eye area frames, that is, based on the operation of connecting the tracking points in the two selected sets of user eye area images, the operation of selecting two sets of user eye area images to set tracking points and connecting the tracking points to obtain the user eye movement path is performed continuously, and the set of user eye area images with the earliest marked timestamp in the next two sets of user eye area images selected and the set of user eye area images with the latest marked timestamp in the last two sets of user eye area images selected are the same set of user eye area images.

[0084] Through the above settings, a specified logic is provided for determining the user's eye sub-movement path in the method in Example 1.

[0085] like Figure 1 As shown, when several groups of user eyeball sub-motion paths are connected to each other, the timestamps of the user eyeball area images of the two end points of each group of user eyeball sub-motion paths are synchronously identified, and the end points of the user eyeball sub-motion paths with the same timestamps are connected to each other to form the user eyeball movement path;

[0086] During the process of acquiring the user's eye movement path, the user's eye moves based on a specified voice prompt, or moves based on a specified dynamic target movement trajectory in an animation;

[0087] After the user's eye movement path is obtained, the nodes on the user's eye movement path are further picked up, and each node is mapped to AutoCAD based on a 1:1 distance ratio between each other. The spline curve command is used to connect each group of mapping points in sequence to obtain a group of curves, and the obtained curves are used to iterate the original user's eye movement path.

[0088] In this embodiment, based on the above settings, the user's eye movement path is further improved, so that the user's eye movement path composed of multiple groups of line segments becomes a continuous curve, so that it is more in line with the movement state of the user's eyes, and at the same time the output result accuracy of the method in Example 1 is improved.

[0089] In summary, during the execution of the method in the above embodiment, by collecting the dynamic image of the user's eye, the user's eye picture frame is further decomposed in the dynamic image of the user's eye, so as to obtain the user's eye movement path based on the tracking point setting, and finally based on the continuous acquisition and difference analysis of the user's eye movement path, an effective auxiliary reference is provided for the determination of the health status of the user's eye, so that medical staff have more effective auxiliary reference data when diagnosing the user's eye problems, so as to make a quick and accurate diagnosis of the user's eye problems.

[0090] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An eye calibration method based on visual image processing technology, characterized in that: include: Step 1: Capture the user's eyeball dynamic image in real time through a high-speed camera, decompose the user's eyeball dynamic image to obtain the user's eyeball picture frame, and mark and store the corresponding source time stamp for each decomposed user eyeball picture frame; Step 2: Select a stored user eyeball picture frame, capture a user eyeball dynamic tracking area in the selected user eyeball picture frame, and further set a tracking point in the captured user eyeball dynamic tracking area; Step 3: Based on the connection lines of the tracking points in the two selected groups of user eyeball picture frames, a plurality of groups of user eyeball sub-motion paths are obtained, and the plurality of groups of user eyeball sub-motion paths are connected to each other to obtain the user eyeball movement path; Step 4: Repeat the above steps to continuously obtain the user's eye movement paths, and analyze the differences in the continuously obtained user's eye movement paths; Step 5: Repeat the above steps to continuously obtain the user's eye movement paths, and analyze the differences in the continuously obtained user's eye movement paths; Step 6: Set the user's eye health status determination logic, and determine whether the user's eyes are healthy based on the differences in the user's eye movement paths and the user's eye health status determination logic.

2. The eye calibration method based on visual image processing technology according to claim 1, characterized in that: When the user eyeball dynamic image captured by the high-speed camera is decomposed into user eyeball picture frames, the number of the decomposed user eyeball picture frames is twice the duration of the user eyeball dynamic image, and each group of decomposed user eyeball picture frames are synchronously identified in the storage stage after marking the corresponding source timestamp in the user eyeball picture frame. The user eyeball area image identified in the user eyeball picture frame is segmented in the user eyeball picture frame, and the segmented user eyeball area image is marked with the source timestamp, and then further stored. When determining the number of frames of the user's eyeball image, the unit of the duration of the user's eyeball dynamic image is defined as seconds.

3. The eye calibration method based on visual image processing technology according to claim 2, characterized in that: The recognition logic of the user's eye area image is expressed as: Two sets of user eye area image determination thresholds are set for the brightness and grayscale value determination of the user eye area, and the brightness and grayscale value of each pixel in the user eye image frame are further calculated: ; Where: is the pixel brightness; is the pixel gray value; , , The three sets of values ​​for the pixel based on the three color channels; Among them, based on the above formula, the brightness and grayscale values ​​of each pixel in the user eyeball picture frame are calculated, and the calculation result is further compared with the corresponding user eyeball area image judgment threshold. The pixels whose brightness and grayscale values ​​meet the user eyeball area image judgment threshold are taken as non-segmentation targets, and the pixels whose brightness and grayscale values ​​do not meet the user eyeball area image judgment threshold are taken as segmentation targets. After the pixels in all user eyeball picture frames are judged as non-segmentation targets and segmentation targets, the user eyeball area image is segmented globally based on the user eyeball picture frame.

4. The eye calibration method based on visual image processing technology according to claim 1, characterized in that: The operation of selecting the stored user eyeball picture frame and capturing the user eyeball dynamic tracking area in the selected user eyeball picture frame, i.e., selecting the user eyeball area image and capturing the user eyeball dynamic tracking area in the user eyeball area image; The interception logic of the user's eye dynamic tracking area in the user's eye area image is expressed as: A set of grayscale determination thresholds for the non-sclera area of ​​the user's eyeball is set, and the grayscale determination thresholds for the non-sclera area of ​​the user's eyeball are compared with pixels in the user's eyeball area image to obtain the non-sclera area of ​​the user's eyeball in the user's eyeball area image, and further obtain the center point of the non-sclera area of ​​the user's eyeball; Analyze and select the similarity of the user's eye area images, and determine the user's eye dynamic tracking area based on the similarity and the center point of the user's eye non-sclera area; Among them, the user's eye dynamic tracking area is a circular area with the center point of the user's eye non-sclera area as the center, and the diameter of the circular area is the product of the maximum inscribed circle diameter of the user's eye area image and the similarity of the user's eye area image.

5. The eye calibration method based on visual image processing technology according to claim 4, characterized in that: The similarity analysis logic of the user's eye area image is expressed as: ; Where: is the similarity between the user's eyeball area image a and the user's eyeball area image b; , is the texture feature vector of the user's eyeball area image a and the user's eyeball area image b; , is the modulus of the two sets of vectors; , is the element value of the user eye area image a and the user eye area image b at the coordinate (i, j), where i and j represent the index of the row and column; , is the element average of the user's eye area image a and the user's eye area image b; , is the weight; in, The value is between 0 and 1. The larger the value, the more similar the eye area images of the two groups of users are. Conversely, the smaller the value, the less similar the eye area images of the two groups of users are. , The sum of is 1, and < .

6. The eye calibration method based on visual image processing technology according to claim 4, characterized in that: When performing the analysis operation on the similarity of the user's eye area images, two groups of user's eye area images are applied, and the two groups of user's eye area images are two groups of user's eye area images that are adjacent based on the marking timestamps; The tracking point set in the user eye dynamic tracking area is a group of pixels with the smallest grayscale value in the user eye dynamic tracking area; When the pixel with the minimum gray value is not unique, a group of pixels with the minimum gray value closest to the center of the user's eye dynamic tracking area is set as the tracking point.

7. The eye calibration method based on visual image processing technology according to claim 1 or 4, characterized in that: Based on the operation of connecting the tracking points in the two selected sets of user eye area frames, that is, based on the operation of connecting the tracking points in the two selected sets of user eye area images, the operation of selecting two sets of user eye area images to set tracking points and connecting the tracking points to obtain the user eye movement path is performed continuously, and the set of user eye area images with the earliest marked timestamp in the next two sets of user eye area images selected and the set of user eye area images with the latest marked timestamp in the last two sets of user eye area images selected are the same set of user eye area images.

8. The eye calibration method based on visual image processing technology according to claim 1, characterized in that: When the plurality of groups of user eyeball sub-motion paths are connected to each other, the timestamps of the user eyeball region images of the endpoints at both ends of each group of user eyeball sub-motion paths are synchronously identified, and the endpoints of the user eyeball sub-motion paths with the same timestamps are connected to each other to form a user eyeball movement path; During the process of acquiring the user's eye movement path, the user's eye completes the eye movement based on a specified voice prompt, or completes the eye movement based on a specified dynamic target movement trajectory in an animation.

9. The eye calibration method based on visual image processing technology according to claim 8, characterized in that: After the user's eye movement path is obtained, nodes on the user's eye movement path are further picked up, and each node is mapped to AutoCAD based on a 1:1 distance ratio between each other. A spline curve command is used to sequentially connect each group of mapping points to obtain a group of curves, and the obtained curves are used to iterate the originally obtained user's eye movement path.

10. The eye calibration method based on visual image processing technology according to claim 1, characterized in that: When analyzing the differences in the eye movement paths of the users, the three groups of eye movement paths of the users that have been acquired the most recently are used for analysis. The three groups of eye movement paths of the users are recorded as x1, x2, and x3. The analysis logic is expressed as follows: The user's eye health status determination logic is expressed as: ; Where: The differences in users’ eye movement paths; , is the number of control points on the eye movement paths of the two groups of users; is the u-th control point on the user's eye movement path x1; is the vth control point on the user's eye movement path x2; , The difference of the eye movement path of the first user and the difference of the eye movement path of the second user; in, represents the magnitude of the distance between two points, Based on x1 and x2, Based on x2 and x3, formula (1) represents the logic of analyzing the difference of the user's eye movement path, and formula (2) represents the logic of judging the user's eye health status. When formula (2) holds, it means that the user's eye health status tends to recover or deteriorate. When formula (2) does not hold, it means that the current judgment result of the user's eye health status is invalid.