Vision detection method and system

Abstract

The invention discloses a vision detection method and system, and belongs to the technical field of eye biological measurement equipment, and the method comprises the steps: obtaining a real-time eye image of a vision detector through an image collection device; identifying a category corresponding to the real-time eye image based on a classification model; if the identified category is not the target category, performing displacement adjustment on the vision detection device based on the identified category; and repeating the steps until the identified category is the target category, and performing vision detection on the eyes of the vision detector based on the adjusted vision detection device. And if the identified category is the target category, performing vision detection on the eyes of the vision detector by using a vision detection device. According to the vision detection method and system, the automation degree, the measurement precision, the safety and the reliability of vision detection are improved, the complexity and the time consumption of automatic adjustment and the interference of eyelids, eyelashes and the like are greatly reduced, and the precision and the speed of focusing adjustment are improved.

Description

Technical Field

[0001] The present application relates to the field of eye biometric measurement equipment, and in particular to a vision detection method and system. Background Art

[0002] The measurement of corneal refraction is a very important part of the diagnosis and treatment of ophthalmic diseases. Currently, existing vision testing equipment requires manual adjustment of the distance from the eye to obtain a clear image to complete the measurement. However, this manual adjustment is highly dependent on the operator's experience and technical level. Different operators may obtain slightly different results, and the measurement accuracy fluctuates.

[0003] Therefore, there is an urgent need for an eye vision detection method and system that can reduce human interference and improve measurement efficiency and accuracy. Summary of the invention

[0004] This application aims to provide a vision detection method and system to reduce human interference and improve measurement efficiency and accuracy.

[0005] In order to achieve the above-mentioned purpose, the technical solution of this application is: In a first aspect, the present application provides a vision detection method, comprising: Providing a vision detection device, wherein the vision detection device is provided with an image acquisition device; Constructing a classification model, wherein the classification model includes multiple categories, each category corresponding to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device; Perform the following detection steps: use the image acquisition device to obtain a real-time eye image of a person undergoing vision test; identify the category corresponding to the real-time eye image based on the classification model; if the identified category is not the target category, perform displacement adjustment on the vision detection device based on the identified category; repeat the above steps of this step until the identified category is the target category, and then perform a vision test on the eyes of the person undergoing vision test based on the adjusted vision detection device; if the identified category is the target category, use the vision detection device to perform a vision test on the eyes of the person undergoing vision test.

[0006] Optionally, the eye image and the real-time eye image both include a pupil area and a halo located in the pupil area, the blurriness of the eye image is proportional to the width of the halo of the eye image, and the blurriness of the real-time eye image is proportional to the width of the halo of the real-time eye image.

[0007] Optionally, among the various categories, the width of the halo of the eye image corresponding to the target category is the smallest.

[0008] Optionally, after using the image acquisition device to acquire the real-time eye image of the person undergoing vision test, and before identifying the category corresponding to the real-time eye image based on the classification model, the method further includes: taking the pupil in the real-time eye image as the center and determining a cropping area of ​​the real-time eye image based on the halo of the real-time eye image; cropping the real-time eye image of the cropping area to obtain a plurality of cropped images; Identifying a category corresponding to the real-time eye image based on the classification model includes: identifying categories corresponding to a plurality of the cropped images based on the classification model.

[0009] Optionally, the category also includes an occlusion category. If the identified category is a non-target category, when the identified category is the occlusion category, a warning message is issued and the above steps are repeated until the identified category is the target category, and then a vision test is performed on the eyes of the vision tester based on the adjusted vision detection device.

[0010] Optionally, the performing displacement adjustment on the vision detection device based on the obtained category comprises: Based on the identified category, the vision detection device is displaced and adjusted using a stepper motor.

[0011] In a second aspect, the present application also provides a vision detection system, comprising: A vision detection device, wherein the vision detection device is provided with an image acquisition device, and the image acquisition device is used to obtain a real-time eye image of a person undergoing vision detection; A classification model building module, used to build a classification model, the classification model includes multiple categories, each category corresponds to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device; the classification model is used to identify the category corresponding to the real-time eye image; A control device is used to perform displacement adjustment on the vision detection device based on the category recognized by the classification module when the category recognized by the classification module is a non-target category; and to control the vision detection device to perform vision detection on the eyes of the vision tester when the category recognized by the classification model is a target category.

[0012] Optionally, the vision detection system further includes: A cropping module, wherein the cropping module is used to determine a cropping area of ​​the real-time eye image based on the halo of the real-time eye image and to crop the real-time eye image in the cropping area to obtain a plurality of cropped images; and the control device identifies categories corresponding to the plurality of cropped images based on the classification model.

[0013] Optionally, the category also includes an occlusion category; if the identified category is a non-target category and is the occlusion category, the control device is further configured to issue a warning message.

[0014] Optionally, the control device comprises: A stepping motor connected to the vision detection device; A control module is connected to the stepper motor and the classification model building module. The control module is used to send a displacement adjustment signal to the stepper motor based on the category identified by the classification module when the category identified by the classification module is a non-target category, and to control the vision detection device to perform a vision detection on the eyes of the vision tester when the category identified by the classification model is a target category; the stepper motor performs displacement adjustment on the vision detection device based on the displacement adjustment signal.

[0015] The vision detection method and system of the present application can achieve automatic focusing of the eyes through the combination of target detection, classification model and automatic displacement adjustment, greatly reducing the complexity and time consumption of automatic adjustment, avoiding human interference, and improving measurement efficiency and accuracy.

[0016] In order to make the above features and advantages of the application more obvious and easy to understand, the following embodiments are specifically cited and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the vision detection method of the present application.

[0018] Figure 2 Schematic diagram of locating the pupil area for this application.

[0019] Figure 3 Schematic diagram of the detection frame of the halo cropped for this application.

[0020] Figure 4 This is a schematic diagram of the modules of the vision detection system of the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose and technical solution of the embodiment of the present application clearer, the technical solution of the embodiment of the present application will be clearly and completely described in conjunction with the drawings of the embodiment of the present application. Obviously, the described embodiment is a part of the embodiment of the present application, not all of the embodiments. Based on the described embodiment of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] In one embodiment, see Figure 1 , the present application provides a vision detection method and system, including: steps S10~S30.

[0023] Step S10: providing a vision detection device, wherein the vision detection device is provided with an image acquisition device.

[0024] Step S20: constructing a classification model, wherein the classification model includes multiple categories, each category corresponding to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device.

[0025] Step S30: using the image acquisition device to obtain a real-time eye image of the person undergoing vision test; identifying the category corresponding to the real-time eye image based on the classification model; if the identified category is not the target category, performing displacement adjustment on the vision detection device based on the identified category; repeating the above steps of this step until the identified category is the target category, performing a vision test on the eyes of the person undergoing vision test based on the adjusted vision detection device; if the identified category is the target category, performing a vision test on the eyes of the person undergoing vision test using the vision detection device.

[0026] The vision detection method and system of the present application can achieve automatic focusing of the eyes through the combination of target detection, classification model and automatic displacement adjustment, greatly reducing the complexity and time consumption of automatic adjustment, avoiding human interference, and improving measurement efficiency and accuracy.

[0027] In step S10, refer to Figure 1 In step S10, a vision detection device is provided, and an image acquisition device is provided on the vision detection device.

[0028] As an example, the image acquisition device is used to obtain a real-time eye image of a person undergoing vision testing, where the real-time eye image includes a pupil area and a halo located in the pupil area.

[0029] As an example, the image acquisition device is also used to obtain and store the eye image of the person undergoing vision test in advance, and the stored eye image is used to construct the classification model in the subsequent step S20. The eye image also includes the pupil area and the halo located in the pupil area.

[0030] In step S20, refer to Figure 1 In step S20, a classification model is constructed, wherein the classification model includes multiple categories, each category corresponding to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device.

[0031] As an example, the blur degree of the eye image is proportional to the width of the halo of the eye image, and the blur degree of the real-time eye image is proportional to the width of the halo of the real-time eye image.

[0032] As an example, the classification model establishes the relationship between the image blur degree based on the width of the halo and the relative displacement of the vision detection device. Specifically, the distance moved by the vision detection device can be divided into 120 categories according to the blur degree of the real-time eye image. When the width of the halo is the smallest, the real-time eye image is the clearest. At this time, the classification label is 60, indicating the best focus position; when the displacement of the vision detection device corresponding to the width of the halo needs to be moved back by 1 step, the classification label is 59, when the displacement of the vision detection device corresponding to the width of the halo needs to be moved back by 2 steps, the classification label is 58, and so on, when the displacement of the vision detection device corresponding to the width of the halo needs to be moved back by 59 steps, the classification label is 1; when the displacement of the vision detection device corresponding to the width of the halo needs to be moved forward by 1 step, the classification label is 61, when the displacement of the vision detection device corresponding to the width of the halo needs to be moved forward by 2 steps, the classification label is 62, and so on, when the halo moves forward by 59 steps, the classification label is 119; when the real-time eye image is blocked, the classification label is 0.

[0033] Furthermore, after the classification model classifies the labels, due to the existence of errors, classification labels 58-62 can be defined as target categories, indicating that the width of the halo is the smallest and the real-time eye image is clearest at this time; then classification labels 1-57, classification labels 63-119 and classification label 0 are non-target categories; among them, classification label 0 is also an occlusion category, indicating that a warning message needs to be issued to enable the vision detection device or the vision tester to make adjustments.

[0034] In step S30, refer to Figure 1 In step S30, after the classification model is constructed, the image acquisition device is used to obtain a real-time eye image of the vision tester; the category corresponding to the real-time eye image is identified based on the classification model; if the identified category is not the target category, the vision detection device is adjusted based on the identified category; the above steps of this step are repeated until the identified category is the target category, and the vision test of the eyes of the vision tester is performed based on the adjusted vision detection device; if the identified category is the target category, the vision test of the eyes of the vision tester is performed using the vision detection device.

[0035] For further information, see Figure 2 , it is necessary to locate the pupil area in the collected real-time eye image, such as Figure 2 As shown, it is a schematic diagram of positioning the pupil area. If the pupil area is detected, the positioning is judged to be successful and a real-time eye image is collected. If it is unsuccessful, the image is re-collected.

[0036] Specifically, a target detection algorithm is used to locate the pupil area in the collected real-time eye image to determine the precise position of the pupil. The target detection algorithm can use the target detection algorithm of the YOLO series to divide the collected image into grids of the same size, process the segmented image in the grid, take the center point of the target object border, and finally use non-maximum suppression processing to locate the precise position of the target.

[0037] As an example, after using the image acquisition device to obtain the real-time eye image of the person with vision tester, before identifying the category corresponding to the real-time eye image based on the classification model, it also includes: taking the pupil in the real-time eye image as the center, determining the cropping area of ​​the real-time eye image based on the halo of the real-time eye image; cropping the real-time eye image of the cropping area to obtain multiple cropped images.

[0038] As an example, see Figure 3 After the pupil area is detected and the precise position of the pupil is determined, the present application uses the YOLO algorithm as an example to detect the position of the halo in the pupil area of ​​the image, draws a halo detection frame, and crops a 224 pixel × 224 pixel square image with the detection frame as the center. The cropped halo detection frame image is as follows Figure 3 As shown in the figure, since there are fewer eyelashes in the pupil area, the influence of eyelids and eyelashes on the classification results is reduced, the accuracy and efficiency of the classification model are improved, and the calculation amount of the model is reduced.

[0039] As an example, identifying the category corresponding to the real-time eye image based on the classification model includes: identifying the categories corresponding to the plurality of cropped images based on the classification model.

[0040] As an example, the cropped image is input into the classification model, and the algorithm outputs the corresponding category according to the blurriness of the cropped image. If the identified category is the target category of the classification label 58-62, the vision detection device can be used directly to perform vision detection on the eyes of the person being tested. If the identified category is a non-target category, the vision detection device drives the vision detection device to adjust its displacement by a stepper motor according to the number of steps indicated by the category. For example, if the identified classification label is 1-57, it means that the vision detection device should move backward 57-1 steps, and if the classification label is 63-119, it means that the vision detection device should move forward 1-57 steps. If the identified category is a non-target category and the classification label is an occlusion category of 0, it means that there is currently an occlusion, such as eyelids, eyelashes or other objects, and normal measurement cannot be performed. A warning message is issued to prompt the vision tester to adjust the eye position or clear the occlusion; the vision detection device is adjusted according to the identified category, and the above steps of this step are repeated. The image acquisition device on the vision detection device is used to capture a real-time image of the eye to determine whether the target detection is successful. If successful, the pupil area is cropped. If unsuccessful, the image is recaptured. After the target detection is successful, the cropped image is input into the classification model for classification until the identified category is the target category and the best focusing effect is obtained. The adjusted vision detection device is used to perform a vision test on the eyes of the vision tester. The movement step size and direction of the vision detection device are determined by the category predicted by the classification model, ensuring a smooth and precise focusing process.

[0041] In another embodiment, see Figure 4 , the present application provides a vision detection system, comprising: A vision testing device 10, wherein the vision testing device 10 is provided with an image acquisition device 101, and the image acquisition device 101 is used to obtain a real-time eye image of a person undergoing vision testing; A classification model building module 20 is used to build a classification model, wherein the classification model includes multiple categories, each category corresponding to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device; the classification model is used to identify the category corresponding to the real-time eye image; The control device 30 is used to perform displacement adjustment on the vision detection device 10 based on the category recognized by the classification module when the category recognized by the classification module is a non-target category; and to control the vision detection device 10 to perform vision detection on the eyes of the vision tester when the category recognized by the classification model is a target category.

[0042] As an example, the vision detection system can be used to perform the following Figure 1 And the steps of the vision detection method mentioned above.

[0043] As an example, please refer to Figure 4The image acquisition device 101 is used to obtain a real-time eye image of a person undergoing vision testing. The real-time eye image includes a pupil area and a halo located in the pupil area.

[0044] As an example, the image acquisition device 101 is also used to obtain and store the eye image of the vision tester in advance, and the stored eye image is used to construct the classification model in the subsequent classification model construction module 20. The eye image also includes the pupil area and the halo located in the pupil area.

[0045] As another example, see Figure 4 , the vision detection system further comprises: A cropping module 40 is used to determine a cropping area of ​​the real-time eye image based on the halo of the real-time eye image with the pupil in the real-time eye image as the center, and crop the real-time eye image in the cropping area to obtain a plurality of cropped images; the classification model identifies categories corresponding to the plurality of cropped images, and the control device 30 identifies categories corresponding to the plurality of cropped images based on the classification model.

[0046] As an example, after determining the cropping area of ​​the real-time eye image based on the halo of the real-time eye image and drawing a halo detection frame, the cropping module 40 is used to crop a square image of 224 pixels×224 pixels with the detection frame as the center.

[0047] As another example, the categories also include an occlusion category.

[0048] As an example, the control device 30 is used to identify the categories corresponding to the plurality of cropped images based on a classification model, and when the identified category is a non-target category and the occlusion category, the control device 30 is used to issue a warning message.

[0049] Wherein, the control device 30 comprises: A stepping motor 301 connected to the vision detection device 30; The control module 302 is connected to the stepper motor 301 and the classification model building module 20. The control module 302 is used to send a displacement adjustment signal to the stepper motor 301 based on the category identified by the classification model built by the classification model building module 20 when the category identified by the classification module is a non-target category, and control the vision detection device 10 to perform a vision test on the eyes of the vision tester when the category identified by the classification model is a target category; the stepper motor 301 performs displacement adjustment on the vision detection device 10 based on the displacement adjustment signal.

[0050] The control device comprises: A stepper motor connected to the vision detection device; A control module is connected to the stepper motor and the classification model building module. The control module is used to send a displacement adjustment signal to the stepper motor based on the category recognized by the classification module when the category recognized by the classification module is a non-target category, and to control the vision detection device to perform a vision detection on the eyes of the vision tester when the category recognized by the classification model is a target category; the stepper motor performs displacement adjustment on the vision detection device based on the displacement adjustment signal.

[0051] The vision detection method and system of the present application have a high degree of automation: by combining target detection with classification models, automatic focus on the eyes is achieved, greatly reducing the complexity and time consumption of automatic adjustment. High measurement accuracy: by first performing pupil detection and cropping the image, the interference of eyelids, eyelashes, etc. is reduced, the complexity and calculation amount of the classification model are also reduced accordingly, and the accuracy and speed of focus adjustment are improved. Using classification models instead of traditional algorithms reduces the number of motor movements and achieves fast focus. Safety and reliability: The detection function of eye occlusion ensures that movement or measurement is stopped in inappropriate situations to prevent the occurrence of erroneous results.

[0052] Although the present application has been disclosed as above with the embodiments, it is not intended to limit the present application. Any person with ordinary knowledge in the technical field can make some changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be determined by the scope of the attached patent application.

Claims

1. A method for visual acuity detection, characterized in that: include: Providing a vision detection device, wherein the vision detection device is provided with an image acquisition device; Constructing a classification model, wherein the classification model includes multiple categories, each category corresponding to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device; The following detection steps are performed: using the image acquisition device to obtain a real-time eye image of the person undergoing vision testing; Based on the classification model, the category corresponding to the real-time eye image is identified; if the identified category is not the target category, the vision detection device is adjusted based on the identified category; the above steps of this step are repeated until the identified category is the target category, and the vision detection of the eyes of the vision tester is performed based on the adjusted vision detection device; if the identified category is the target category, the vision detection device is used to perform the vision detection of the eyes of the vision tester.

2. The vision detection method according to claim 1, characterized in that: The eye image and the real-time eye image both include a pupil area and a halo located in the pupil area, the blurriness of the eye image is proportional to the width of the halo of the eye image, and the blurriness of the real-time eye image is proportional to the width of the halo of the real-time eye image.

3. The vision detection method according to claim 2, characterized in that: Among the various categories, the width of the halo of the eye image corresponding to the target category is the smallest.

4. The vision detection method according to claim 2, characterized in that: After using the image acquisition device to acquire the real-time eye image of the vision tester, and before identifying the category corresponding to the real-time eye image based on the classification model, the method further includes: taking the pupil in the real-time eye image as the center and determining the cropping area of ​​the real-time eye image based on the halo of the real-time eye image; cropping the real-time eye image of the cropping area to obtain a plurality of cropped images; Identifying a category corresponding to the real-time eye image based on the classification model includes: identifying categories corresponding to a plurality of the cropped images based on the classification model.

5. The vision detection method according to claim 1, characterized in that: The categories also include occlusion categories. If the identified category is a non-target category, when the identified category is the occlusion category, a warning message is issued and the above steps are repeated until the identified category is the target category. The vision tester's eyes are then subjected to a vision test based on the adjusted vision test device.

6. The vision detection method according to any one of claims 1 to 5, characterized in that: The step of adjusting the displacement of the vision detection device based on the obtained category comprises: Based on the identified category, the vision detection device is displaced and adjusted using a stepper motor.

7. A vision detection system, characterized in that: include: A vision detection device, wherein the vision detection device is provided with an image acquisition device, and the image acquisition device is used to obtain a real-time eye image of a person undergoing vision detection; A classification model building module, used to build a classification model, the classification model includes multiple categories, each category corresponds to a different correspondence between the blur degree of the eye image and the displacement of the vision detection device; the classification model is used to identify the category corresponding to the real-time eye image; A control device is used to adjust the displacement of the vision detection device based on the category recognized by the classification module when the category recognized by the classification module is a non-target category; and to control the vision detection device to perform vision detection on the eyes of the vision tester when the category recognized by the classification model is a target category.

8. The vision detection system according to claim 7, characterized in that: Also includes: a cropping module, the cropping module being used to determine a cropping area of ​​the real-time eye image based on a halo of the real-time eye image with the pupil in the real-time eye image as the center, and to crop the real-time eye image in the cropping area to obtain a plurality of cropped images; The control device identifies categories corresponding to the plurality of cropped images based on the classification model.

9. The vision detection system according to claim 7, characterized in that: The categories also include occlusion categories; if the identified category is a non-target category and is the occlusion category, the control device is further used to issue a warning message.

10. The vision detection system according to any one of claims 7 to 9, characterized in that: The control device comprises: A stepping motor connected to the vision detection device; A control module is connected to the stepper motor and the classification model building module. The control module is used to send a displacement adjustment signal to the stepper motor based on the category identified by the classification module when the category identified by the classification module is a non-target category, and to control the vision detection device to perform a vision detection on the eyes of the vision tester when the category identified by the classification model is a target category; the stepper motor performs displacement adjustment on the vision detection device based on the displacement adjustment signal.

Images