CN109919896A - A kind of measurement method of corneal diameter - Google Patents
A kind of measurement method of corneal diameter Download PDFInfo
- Publication number
- CN109919896A CN109919896A CN201711308312.9A CN201711308312A CN109919896A CN 109919896 A CN109919896 A CN 109919896A CN 201711308312 A CN201711308312 A CN 201711308312A CN 109919896 A CN109919896 A CN 109919896A
- Authority
- CN
- China
- Prior art keywords
- image
- scale
- corneal
- boundary
- central area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The present invention relates to a kind of measurement methods of corneal diameter, comprising: obtains the image for carrying cornea；It determines corneal center region, determining corneal center region is known as the first central area；It is respectively that central area intercepts two images as corneal boundary image using the second central area and third central area；Second central area and third central area are that the first central area is offset to both sides the region after pre-determined distance respectively；Two corneal boundary reference lines are determined on two corneal boundary images respectively；Corneal diameter length is obtained according to two corneal boundary reference lines.The application measurement accuracy is high, measurement method is simple and fast and at low cost.
Description
Technical field
The present invention relates to a kind of measurement methods of corneal diameter, belong to field of image processing.
Background technique
Cornea is eyes front end layer of transparent film, and measure corneal diameter has important work in the diagnosis of ophthalmology disease
With.For example, infant development speed is quickly, measurement infant's corneal diameter variation not only understands its growth and development situation, but also
There is valuable help to the diagnosis of congenital glaucoma.Newborn's corneal diameter is greater than 11mm, and infant is greater than 12mm very within 1 years old
There may be intraocular pressure raisings.It, generally can be in 12mm or less after being greater than 3 years old.The children of any age level, corneal diameter are greater than
13mm and the obvious asymmetry of cornea of both eyes diameter belong to exception.Therefore measurement infant's corneal diameter, which just seems, especially closes
Key.
The prior art measures infant's corneal diameter there are mainly two types of mode, and a kind of mode is to pass through mechanical equipment
Carry out, but mechanical equipment low measurement accuracy, generally in 0.5mm hereinafter, and since infant will not cooperate, can in measurement
Mechanical equipment can be caused to damage the eyes of infant, in order to accurately measure, infant can be injected in measurement
Tranquillizer applies general anaesthetic, and not only expense is high, these drugs can also have an impact infant's body；Another way is
Using automatic measuring equipment, for example, optometry curvature instrument, OrbscanIIz eye anterior segment analysis system, Optic coherence biometry and
Corneal topography etc., but these automatic measuring equipments are expensive, and operation is more complicated, and portability is poor, while not being suitable for again
The measurement of infant's cornea transverse diameter.
Summary of the invention
The invention proposes a kind of measurement method of corneal diameter, the corneal diameter precision that this method not only measures is high, and
And measurement method is simple and fast, at low cost.
The embodiment of the invention provides a kind of measurement methods of corneal diameter, comprising:
Obtain the image for carrying cornea；
It determines corneal center region, determining corneal center region is known as the first central area；
It is respectively that central area intercepts two images as corneal boundary figure using the second central area and third central area
Picture；Second central area and third central area are that the first central area is offset to both sides the area after pre-determined distance respectively
Domain；
Two corneal boundary reference lines are determined on two corneal boundary images respectively；
Corneal diameter length is obtained according to two corneal boundary reference lines.
Wherein, the image for carrying cornea is obtained, specifically:
Scale is placed in immediately ahead of eyes；
Image collection module obtains the image for carrying scale, will carry the image of scale as the image for carrying cornea.
Wherein, the scale is made of handle and framework；The area of the framework is greater than preset value.
Further, framework centre mark is carried in the framework.
Further, the framework is rectangle.
Further, the framework outer minister 20mm to 60mm, wide 20mm to 50mm；Framework inner spare area domain
Long 15mm to 45mm, wide 15mm to 40mm.
Still further, the outer minister 43mm, wide 28mm of the framework；The long 35mm in framework inner spare area domain, it is wide
20mm。
Further, the scale is integrated or separately made of stainless steel.
Further, the rectangle framework is in green.
Further, the centre mark is the crosspoint that two straight lines are diagonally arranged in the framework.
Still further, the centre mark is the crosspoint that two red lines are diagonally arranged in the framework.
Further, before determining corneal center region, further includes:
Image included by scale is extracted on the image for carrying scale, and the image extracted is known as scale image.
Wherein, image included by scale is extracted on the image for carrying scale, specifically:
The color space for carrying the image of scale is transformed into the space LAB by rgb space；
Scale image is extracted on the image of the carrying scale in the space LAB based on histogram.
Further, the rectangle framework of the scale is in green；
Correspondingly, scale image is extracted on the image for carrying scale based on histogram, comprising:
Scale image is extracted based on A channel histogram in the space LAB on the image for carrying scale.
Further, further includes:
The scale map that verification is extracted seems no correct, if not, based on Kmeans clustering algorithm in the image for carrying scale
Upper extraction scale image.
Further, on the image for carrying scale after extraction scale image, further includes:
Scale image is normalized.
Wherein, scale image is normalized, specifically:
Scale image is zoomed into specified size；
Point centered on scale centre after scaling intercepts the image of specified size as the scale image after normalization.
Wherein it is determined that corneal center region, specifically:
Centered on the framework centre mark, the graphics field of specified pixel is intercepted, the graphics field of interception is known as
First central area.
Further, further includes:
Framework centre mark is corrected in first central area.
Wherein, two corneal boundary reference lines are determined on two corneal boundary images respectively, comprising:
Respectively by two corneal boundary image binaryzations, two corneal boundary bianry images are obtained；
The boundary for extracting two bianry images respectively, using the vertical tangent line in the outside on each boundary as corneal boundary benchmark
Line.
Further, further includes:
Corresponding corneal boundary image is adjusted separately according to two corneal boundary reference lines；
Two corneal boundary reference lines are determined again according to corneal boundary image adjusted.
Wherein, corneal diameter length is obtained according to two corneal boundary reference lines, specifically:
The coordinate that two corneal boundary reference lines are respectively relative to respective corneal boundary image is converted in cornea image
In coordinate；
According to the lateral shift value of two corneal boundary reference lines coordinate in cornea image, the pixel of corneal diameter is obtained
Span value；
Corneal diameter length is obtained according to the pixel span value of corneal diameter.
Wherein, cornea length is obtained according to the pixel span value of corneal diameter, specifically:
If the physical length of corneal diameter is Lcornea, the pixel span value of corneal diameter is Pcornea, the green frame of cornea image
Pixel quantity in range is Nimage, the inside casing area of scale material object is Aimage, then LcorneaFor
The embodiment of the invention also provides a kind of measuring systems of corneal diameter, comprising: image acquiring device, positioning mould
Block, image interception module, determining module and computing module；
Image acquiring device, for obtaining the image for carrying cornea；
Determining corneal center region is known as the first central area for determining corneal center region by locating module；
Image interception module, for being respectively that central area intercepts two figures with the second central area and third central area
As being used as corneal boundary image；Second central area and third central area are that the first central area is offset to both sides respectively
Region after pre-determined distance；
Determining module, for determining two cornea sides on two corneal boundary images of image interception module intercepts respectively
Boundary's reference line；
Computing module, two corneal boundary benchmark line computation corneal diameter length for being determined according to determining module.
Wherein, described image acquisition device includes image collection module and scale；
The scale is placed in immediately ahead of eyes；
Described image obtains module and is used to obtain the image for carrying scale, will carry the image of scale as carrying cornea
Image；
The scale is made of handle and framework；The area of the framework is greater than preset value.
Further, framework centre mark is carried in the framework.
Further, the embodiment of the invention also includes extraction modules, the carrying scale for obtaining in image acquiring device
Image on extract scale included by image, the image extracted is known as scale image.
Wherein, the extraction module includes color conversion unit and the first extraction unit；
Color conversion unit, the color space of the image of the carrying scale for obtaining image acquiring device is by RGB sky
Between be transformed into the space LAB；
First extraction unit, for extracting scale image on the image of the carrying scale in the space LAB based on histogram.
Further, the rectangle framework of the scale is in green；
Correspondingly, first extraction unit is specifically used for straight based on A channel in the space LAB on the image for carrying scale
Side's figure extracts scale image.
Further, the extraction module further includes verification unit and the second extraction unit；
Verification unit, the scale map for verifying the extraction of the first extraction unit seem it is no correct, if not, triggering second mentions
Take unit；
Second extraction unit, for extracting scale image on the image for carrying scale based on Kmeans clustering algorithm.
It further, further include normalization module, for scale image to be normalized.
Wherein, the normalization module includes unit for scaling and normalization unit；
The unit for scaling, for scale image to be zoomed to specified size；
The normalization unit intercepts specified for the point centered on the scale centre after unit for scaling scaling
The image of size is as the scale image after normalization.
Wherein, the locating module is specifically used for centered on the framework centre mark, intercepts the figure of specified pixel
The graphics field of interception is known as the first central area by region.
It further, further include correction module, for correcting framework centre mark in first central area.
Wherein, the determining module includes image processing unit and Boundary Extraction unit；
Described image processing unit, for obtaining two corneal boundaries two respectively by two corneal boundary image binaryzations
It is worth image；
Boundary Extraction unit, for extracting the boundary of two bianry images respectively, by the vertical tangent line in the outside on each boundary
As corneal boundary reference line.
Further, the determining module further includes adjustment unit, two bases for being determined according to Boundary Extraction unit
Directrix adjusts separately corresponding corneal boundary image, then triggers described image processing unit and Boundary Extraction unit again.
Wherein, the computing module includes coordinate transformation unit, the first computing unit and the second computing unit；
The coordinate transformation unit, for two corneal boundary reference lines to be respectively relative to respective corneal boundary image
Coordinate is converted to the coordinate in cornea image；
First computing unit, for the lateral shift according to two corneal boundary reference line coordinates in cornea image
Value, calculates the pixel span value of corneal diameter；
Second computing unit, based on the pixel span value of the corneal diameter by being calculated according to first computing unit
Calculate corneal diameter length.
If the physical length of corneal diameter is Lcornea, the pixel span value of corneal diameter is Pcornea, the green frame of cornea image
Pixel quantity in range is Nimage, the inside casing area of scale material object is Aimage, then second computing unit calculates
LcorneaFor
The beneficial effects of the present invention are:
After the embodiment of the present invention obtains the image for carrying cornea, corneal center region is determined, then according to corneal center area
Domain obtains corneal boundary image, finally utilizes the corneal boundary reference line determined on corneal boundary image to obtain corneal diameter long
It spends, for mechanical measuring means compared with the existing technology, not only measurement accuracy is high, but also measurement method is simple and fast, is not easy
It is influenced by the activity of infant, due to being not required to tranquillizer or anaesthetic, infant will not both be damaged, and eliminate medicine
Object cost；For method for automatic measurement compared with the existing technology, not only measurement cost is cheap, and measurement accuracy is high, and measures
Operation is simple；
Further, then the present invention obtains the image for carrying scale, without connecing by scale is arranged immediately ahead of eyes
It contacts that eyes can measure, and makes the image of the carrying cornea obtained more accurate；
Further, scale is made of handle and framework, easy to carry；
Further, it is determined that scale image is extracted on the image for carrying scale, so that subsequent before corneal center region
Processing is more quick, accurate, also improves the precision of measurement；
Further, the embodiment of the present invention carry scale image on extract scale image after also to scale image into
Row normalized handles scale image under same coordinate standard, reduces the complexity of subsequent processing.
Detailed description of the invention
Fig. 1 is a kind of measurement method flow diagram of corneal diameter provided by the invention；
Fig. 2 is scale schematic diagram provided by the invention；
Fig. 3 is a kind of measuring system structural schematic diagram of corneal diameter provided by the invention.
Specific embodiment
In order to keep the purpose of the present invention, technical solution and advantage etc. more explicit, concrete instance is enumerated herein
And it is further illustrated with refer-ence to the attached drawings.
Referring to Fig. 1, the embodiment of the invention provides a kind of measurement methods of corneal diameter, comprising:
101, the image for carrying cornea is obtained；
Fig. 2 is a kind of scale schematic diagram provided in an embodiment of the present invention, can more accurately be obtained and be taken by the scale
Image with cornea.The scale is made of handle and framework, and the area of framework is greater than preset value, which can be according to being clapped
The eye socket area for the person of taking the photograph determines.Framework is rectangle in the embodiment of the present invention, which can integrated or separately be made of stainless steel.
Further, in order to which the cornea image of acquisition is more accurate, framework centre mark can be set in framework, make frame
Body centre mark and the alignment of eye pupil center, framework centre mark can be demarcated by the cornerwise crosspoint of framework,
It can be demarcated by the crosspoint of long side center line and broadside center line, the embodiment of the present invention is specific to framework centre mark
Calibration mode is with no restriction.
Scale inside casing used in the embodiment of the present invention is long 43mm, money having a size of long 35mm, wide 20mm, frame size
28mm, interior outline border spacing be 4mm, grip size be long 50mm, money 4mm, scale with a thickness of 1mm, the rectangle framework of scale
Peripheral radius of corner is 1mm.
Specifically, the image for carrying cornea can obtain in the following way:
Scale is placed in immediately ahead of eyes；
Image collection module obtains the image for carrying scale, will carry the image of scale as the image for carrying cornea.
In actual use, scale can be placed at the eye socket immediately ahead of the person's of being taken eyes, when placement, make scale
Frame plane face at eye level is parallel as far as possible and is bonded, and by the framework centre mark of scale and the person's of being taken eye pupil center pair
Together, then the scale framework on the image collection module face person's of being taken eye socket is obtained to the image for carrying scale, optimal, figure
As obtaining module and scale framework distance about 10cm~15cm.In the embodiment of the present invention image collection module can be camera or
Mobile phone etc..
Further, in order to keep the corneal diameter precision of measurement higher, the embodiment of the present invention is determining corneal center region
Before, further includes:
Image included by scale is extracted on the image for carrying scale, and the image extracted is known as scale image.
Specifically, scale image is extracted on the image for carrying scale, are as follows:
The color space for carrying the image of scale is transformed into the space LAB by rgb space；
Scale image is extracted on the image of the carrying scale in the space LAB based on histogram.
Compared with RGB color, LAB is a kind of and device-independent color system, and a kind of based on physiological characteristic
Color system.That is, LAB is the visual response for describing people with method for digitizing.L points in LAB color space
The brightness for indicating pixel is measured, value range is [0,100], is indicated from black to pure white；A indicates the model from red to green
It encloses, value range is [127, -128]；B indicates the range from yellow to blue, and value range is [127, -128].
Histogram is a kind of common image processing method, it describes the probability distribution of gray value of image, wherein face
Color can regard the joint probability distribution of the gray value in different channels as, can also intuitively be considered the color that human eye is perceived
Distribution.
Further, in order to more convenient based on histogram extraction scale image, by the length of scale in the embodiment of the present invention
Square frame is set as green, correspondingly, scale image is extracted on the image for carrying scale based on histogram are as follows:
Scale image is extracted based on A channel histogram in the space LAB on the image for carrying scale.
Since the rectangle framework of scale is in green, lack haematochrome, and the negative axis directions pair of the A channel in the space LAB
It is quasi- red, scale image zooming-out is carried out to the A channel in the space LAB, the accuracy of scale image zooming-out can greatly be improved
And stability.
In practical application, due to light environment complexity, it is possible that mistake, in order to avoid the influence to subsequent processing,
The embodiment of the invention also includes:
The scale map that verification is extracted seems no correct, if not, based on Kmeans clustering algorithm in the image for carrying scale
Upper extraction scale image.
The scale image accuracy extracted based on Kmeans clustering algorithm is relatively high, but speed is very slow, is being based on
After Kmeans clustering algorithm extracts scale image to the A channel in the space LAB, the scale image of extraction can also be verified again
It is whether correct, if verification is incorrect, Kmeans clustering algorithm can be again based on, scale is extracted to the channel AB in the space LAB
Image, until the scale image verification success of extraction.
Further, in order to reduce the complexity of measurement method, scale map picture can be made to carry out under same coordinate standard
Processing, therefore the embodiment of the invention also includes:
Scale image is normalized.
Specifically, scale image is normalized, are as follows:
Scale image is zoomed into specified size；
Point centered on scale centre after scaling intercepts the image of specified size as the scale image after normalization.
For example, scale image is carried out undistorted scaling, keep the area in its scale region small as far as possible, i.e., fixed length and width
The minimum of ratio is outer to cut rectangle, and then point centered on the scale picture centre after scaling, the rectangle that interception is one 700 × 400 are made
For the scale image after normalization.
102, it determines corneal center region, determining corneal center region is known as the first central area；
Specifically, centered on framework centre mark, the graphics field of specified pixel is intercepted, the graphics field of interception is claimed
For the first central area.
For example, interception size is the square figure of 50 pixels centered on the rectangle framework diagonal line crosspoint of scale
Shape region, using the square-shaped patterns region having a size of 50 pixels as the first central area.
After determining the first central area, the rectangle framework centre mark of scale may be deviated, therefore the present invention
Embodiment is after determining the first central area, further includes:
Framework centre mark is corrected in the first central area.
Specifically, canny operator and straight line hough transformation be can use, correct rectangle in square-shaped patterns region
The position in framework diagonal line crosspoint.Canny operator and straight line hough transformation belong to the common knowledge of those skilled in the art,
Details are not described herein for the embodiment of the present invention.
It 103, is respectively that central area intercepts two images as cornea side using the second central area and third central area
Boundary's image；Second central area and third central area are that the first central area is offset to both sides the area after pre-determined distance respectively
Domain；
In order to keep subsequent calculating simpler, truncated picture shape is rectangle in the embodiment of the present invention, but the present invention is real
It applies example and is not limited to rectangle, as long as the image shape that can reach the object of the invention is ok.
104, two corneal boundary reference lines are determined on two corneal boundary images respectively；
Specifically, two corneal boundary reference lines are determined on two corneal boundary images respectively, comprising:
Respectively by two corneal boundary image binaryzations, two corneal boundary bianry images are obtained；
The boundary for extracting two bianry images respectively, using the vertical tangent line in the outside on each boundary as corneal boundary benchmark
Line.
More preferably, it is bianry image that maximum variance between clusters, which can be used, by corneal boundary image binaryzation, then to this
Bianry image extracts boundary, finally to the vertical tangent line of the standardized item in the outside of arc-shaped side circle, using the vertical tangent line as corneal boundary
Reference line.Maximum between-cluster variance, which refers to, is divided into two classes for entire data using a threshold value, if the variance between two classes
Maximum, then this threshold value is exactly optimal threshold value.It the use of maximum variance between clusters by image binaryzation is those skilled in the art
The common knowledge of member, details are not described herein for the embodiment of the present invention.
Since corneal boundary image is there may be the unbalanced phenomenon of monochrome pixels accounting, this phenomenon will cause image side
Boundary's detection inaccuracy, therefore in order to avoid this phenomenon, the embodiment of the invention also includes:
Corresponding corneal boundary image is adjusted separately according to two corneal boundary reference lines；
Two corneal boundary reference lines are determined again according to corneal boundary image adjusted.
Specifically, corresponding corneal boundary image can be cut according to determining reference line, so that corneal boundary
Image corresponds to reference line bilateral symmetry along it, and keeps the area of the corneal boundary image cut as maximum as possible, after cutting
Corneal boundary image left and right monochrome pixels accounting it is more balanced, then two corneal boundary images after cutting are held again
Row can make measurement result more accurate and steady the step of determining two border level lines on two corneal boundary images
It is fixed.
105, corneal diameter length is obtained according to two corneal boundary reference lines.
Specifically, corneal diameter length is obtained according to two corneal boundary reference lines, are as follows:
The coordinate that two corneal boundary reference lines are respectively relative to respective corneal boundary image is converted in cornea image
In coordinate；
According to the lateral shift value of two corneal boundary reference lines coordinate in cornea image, the pixel of corneal diameter is obtained
Span value；
Corneal diameter length is obtained according to the pixel span value of corneal diameter.
Specifically, if the physical length of corneal diameter is Lcornea, the pixel span value of corneal diameter is Pcornea, cornea figure
As the pixel quantity within the scope of green frame is Nimage, the inside casing area of scale material object is Aimage, then second computing unit calculates
LcorneaFor
In the embodiment of the present invention, step 102 to 105 be can integrate in a server, and step 101, which obtains, carries cornea
Image after, can be uploaded to by network in the server, not have to that specific software is installed, by the server to carrying cornea
Image identified, and measure corneal diameter, can also be by step 102 to 105 integrated APP processing, can basis
Specific to need to be arranged, with no restriction to implementation, therefore the measurement method of corneal diameter provided by the invention is non-by the present invention
It is often easy universal.
After the embodiment of the present invention obtains the image for carrying cornea, corneal center region is determined, then according to corneal center area
Domain obtains corneal boundary image, finally utilizes the corneal boundary reference line determined on corneal boundary image to obtain corneal diameter long
It spends, for mechanical measuring means compared with the existing technology, not only measurement accuracy is high, but also measurement method is simple and fast, is not easy
It is influenced by the activity of infant, due to being not required to tranquillizer or anaesthetic, infant will not both be damaged, and eliminate medicine
Object cost；For method for automatic measurement compared with the existing technology, not only measurement cost is cheap, and measurement accuracy is high, and measures
Operation is simple；Further, then the present invention obtains the image for carrying scale by scale is arranged immediately ahead of eyes,
It can measure without touching eyes i.e., and make the image of the carrying cornea obtained more accurate；Further, scale is by hand
Handle and framework composition, it is easy to carry；Further, it is determined that extracting mark on the image for carrying scale before corneal center region
Ruler image also improves the precision of measurement so that subsequent processing is more quick, accurate；Further, the embodiment of the present invention is being taken
Scale image is extracted on image with scale also scale image is normalized later, make scale image in same coordinate
It is handled under standard, reduces the complexity of subsequent processing.
Referring to Fig. 3, the embodiment of the invention also provides a kind of measuring systems of corneal diameter, comprising: image acquiring device
31, locating module 32, image interception module 33, determining module 34 and computing module 35；
Image acquiring device 31, for obtaining the image for carrying cornea；
Wherein, image acquiring device 31 includes image collection module and scale；
The scale is placed in immediately ahead of eyes；Described image obtains module and is used to obtain the image for carrying scale, will take
Image with scale is as the image for carrying cornea；
The scale is made of handle and framework；The area of the framework is greater than preset value.
Further, framework is rectangle, and framework centre mark is carried in the rectangle framework.
In actual use, scale can be placed at the eye socket immediately ahead of the person's of being taken eyes, when placement, make scale
Frame plane face at eye level is parallel as far as possible and is bonded, and by the framework centre mark of scale and the person's of being taken eye pupil center pair
Together, then the scale framework on the image collection module face person's of being taken eye socket is obtained to the image for carrying scale, optimal, figure
As obtaining module and scale framework distance about 10cm~15cm.In the embodiment of the present invention image collection module can be camera or
Mobile phone etc..
Rectangle framework centre mark can be demarcated by the cornerwise crosspoint of framework, can also pass through long side center
The crosspoint of line and broadside center line is demarcated, and the embodiment of the present invention do not limit the specific calibration mode of framework centre mark
System.
Further, the embodiment of the invention also includes extraction modules, the carrying scale for obtaining in image acquiring device
Image on extract scale included by image, the image extracted is known as scale image.
Specifically, extraction module includes color conversion unit and the first extraction unit；
Color conversion unit, the color space of the image of the carrying scale for obtaining image acquiring device is by RGB sky
Between be transformed into the space LAB；
First extraction unit, for extracting scale image on the image of the carrying scale in the space LAB based on histogram.
Specifically, when the rectangle framework of scale is in green, correspondingly, the first extraction unit is specifically used for carrying mark
Scale image is extracted based on A channel histogram in the space LAB on the image of ruler.
Since the rectangle framework of scale is in green, lack haematochrome, and the negative axis directions pair of the A channel in the space LAB
It is quasi- red, scale image zooming-out is carried out to the A channel in the space LAB, the accuracy that item eats image zooming-out can greatly be improved
And stability.
In practical application, due to light environment complexity, it is possible that mistake, in order to avoid the influence to subsequent processing,
Therefore extraction module further includes verification unit and the second extraction unit in the embodiment of the present invention；
Verification unit, the scale map for verifying the extraction of the first extraction unit seem it is no correct, if not, triggering second mentions
Take unit；
Second extraction unit, for extracting scale image on the image for carrying scale based on Kmeans clustering algorithm.
The scale image accuracy extracted based on Kmeans clustering algorithm is relatively high, but speed is very slow, extracts second
After unit extracts scale image to the A channel in the space LAB based on Kmeans clustering algorithm, it is single that verification can also be triggered again
Member come the scale map that verifies extraction seem it is no correct, if verification is incorrect, extraction module further includes third extraction unit, is used
In based on Kmeans clustering algorithm in the space LAB the channel AB extract scale image, until extraction scale image verification at
Function.
Further, in order to reduce the complexity of measurement method, scale map picture can be made to carry out under same coordinate standard
Processing, therefore the embodiment of the invention also includes normalization modules, for scale image to be normalized.
Specifically, normalization module includes unit for scaling and normalization unit；
Unit for scaling, for scale image to be zoomed to specified size；
Normalization unit, for by unit for scaling scale after scale centre centered on point, intercept the figure of specified size
As the scale image after normalization.
For example, scale image is carried out undistorted scaling by unit for scaling, keep the area in its scale region small as far as possible, i.e.,
The minimum of fixed aspect ratio is outer to cut rectangle, then point centered on scale picture centre of the normalization unit after scaling, interception
One 700 × 400 rectangle is as the scale image after normalization.
Determining corneal center region is known as the first central area for determining corneal center region by locating module 32；
Specifically, locating module 32 intercepts the graphics field of specified pixel centered on framework centre mark, by interception
Graphics field is known as the first central area.
For example, locating module 32, centered on the rectangle framework diagonal line crosspoint of scale, interception size is 50 pixels
Square-shaped patterns region, using the square-shaped patterns region having a size of 50 pixels as the first central area.
After determining the first central area, the rectangle framework centre mark of scale may be deviated, therefore the present invention
Embodiment further includes correction module, for correcting framework centre mark in the first central area.
Specifically, canny operator and straight line hough transformation be can use, correct rectangle in square-shaped patterns region
The position in framework diagonal line crosspoint.
Image interception module 33, for being respectively that central area intercepts two with the second central area and third central area
Image is as corneal boundary image；Second central area and third central area are offset to both sides pre- respectively for the first central area
If the region after distance；
Determining module 34, for determining two cornea sides on two corneal boundary images that image interception module 33 intercepts
Boundary's reference line；
Specifically, it is determined that module 34 includes image processing unit and Boundary Extraction unit；
Wherein, image processing unit, for obtaining two corneal boundaries two respectively by two corneal boundary image binaryzations
It is worth image；
In practical application, maximum variance between clusters are can be used in image processing unit, are by corneal boundary image binaryzation
Bianry image.
Boundary Extraction unit, for extracting the boundary of two bianry images respectively, by the vertical tangent line in the outside on each boundary
As corneal boundary reference line.
It should be noted that there may be the unbalanced phenomenon of monochrome pixels accounting, this phenomenon meetings for corneal boundary image
Image boundary detection inaccuracy is caused, therefore in order to avoid this phenomenon, determining module further includes adjustment in the embodiment of the present invention
Unit, two reference lines for being determined according to Boundary Extraction unit adjust separately corresponding corneal boundary image, then again
Image processing unit and Boundary Extraction unit are triggered, i.e., determines the reference line of cornea transverse diameter again.
Computing module 35, two corneal boundary benchmark line computation corneal diameters for being determined according to determining module 34 are long
Degree.
Specifically, computing module 35 includes coordinate transformation unit, the first computing unit and the second computing unit；
Wherein, coordinate transformation unit, for two corneal boundary reference lines to be respectively relative to respective corneal boundary image
Coordinate be converted to the coordinate in cornea image；
First computing unit, for the lateral shift value according to two corneal boundary reference lines coordinate in cornea image,
Calculate the pixel span value of corneal diameter；
The pixel span value of second computing unit, the corneal diameter for being calculated according to first computing unit calculates angle
Film diameter length.
If the physical length of corneal diameter is Lcornea, the pixel span value of corneal diameter is Pcornea, the green frame of cornea image
Pixel quantity in range is Nimage, the inside casing area of scale material object is Aimage, then second computing unit calculates
LcorneaFor
After the embodiment of the present invention obtains the image for carrying cornea, corneal center region is determined, then according to corneal center area
Domain obtains corneal boundary image, finally utilizes the corneal boundary reference line determined on corneal boundary image to obtain corneal diameter long
It spends, for mechanical measuring means compared with the existing technology, not only measurement accuracy is high, but also measurement method is simple and fast, is not easy
It is influenced by the activity of infant, due to being not required to tranquillizer or anaesthetic, infant will not both be damaged, and eliminate medicine
Object cost；For method for automatic measurement compared with the existing technology, not only measurement cost is cheap, and measurement accuracy is high, and measures
Operation is simple；Further, then the present invention obtains the image for carrying scale by scale is arranged immediately ahead of eyes,
It can measure without touching eyes i.e., and make the image of the carrying cornea obtained more accurate；Further, scale is by hand
Handle and framework composition, it is easy to carry；Further, it is determined that extracting mark on the image for carrying scale before corneal center region
Ruler image also improves the precision of measurement so that subsequent processing is more quick, accurate；Further, the embodiment of the present invention is being taken
Scale image is extracted on image with scale also scale image is normalized later, make scale image in same coordinate
It is handled under standard, reduces the complexity of subsequent processing.
The above is only several embodiments of the application, not does any type of limitation to the application, although this Shen
Please disclosed as above with preferred embodiment, however not to limit the application, any person skilled in the art is not taking off
In the range of technical scheme, a little variation or modification are made using the technology contents of the disclosure above and is equal to
Case study on implementation is imitated, is belonged in technical proposal scope.
Claims (10)
1. a kind of measurement method of corneal diameter characterized by comprising
Obtain the image for carrying cornea；
It determines corneal center region, determining corneal center region is known as the first central area；
It is respectively that central area intercepts two images as corneal boundary image using the second central area and third central area；Institute
It states the second central area and third central area is that the first central area is offset to both sides the region after pre-determined distance respectively；
Two corneal boundary reference lines are determined on two corneal boundary images respectively；
Corneal diameter length is obtained according to two corneal boundary reference lines.
2. the method according to claim 1, wherein obtain carry cornea image, specifically:
Scale is placed in immediately ahead of eyes；
Image collection module obtains the image for carrying scale, will carry the image of scale as the image for carrying cornea.
3. according to the method described in claim 2, it is characterized in that, the scale is made of handle and framework；
The area of the framework is greater than preset value；
Framework centre mark is carried in the framework.
4. according to the method described in claim 3, it is characterized in that, the framework is rectangle；
The scale is integrated or separately made of stainless steel；
The framework is in green.
5. according to the method described in claim 3, it is characterized in that, before determining corneal center region, further includes:
Image included by scale is extracted on the image for carrying scale, and the image extracted is known as scale image.
6. according to the method described in claim 5, it is characterized in that, extracting figure included by scale on the image for carrying scale
Picture, specifically:
The color space for carrying the image of scale is transformed into the space LAB by rgb space；
Scale image is extracted on the image of the carrying scale in the space LAB based on histogram, specifically:
The framework is in green, correspondingly, extracts scale based on A channel histogram in the space LAB on the image for carrying scale
Image；
The scale map that verification is extracted seems no correct, if not, above being mentioned based on Kmeans clustering algorithm in the image for carrying scale
Take scale image.
7. according to the method described in claim 5, it is characterized in that, carry scale image on extract scale image after,
Further include:
Scale image is normalized, specifically:
Scale image is zoomed into specified size；
Point centered on scale centre after scaling intercepts the image of specified size as the scale image after normalization.
8. according to the method described in claim 3, it is characterized in that, determine corneal center region, specifically:
Centered on the framework centre mark, the graphics field of specified pixel is intercepted, the graphics field of interception is known as first
Central area；
Framework centre mark is corrected in first central area.
9. the method according to claim 1, wherein two corneas are determined on two corneal boundary images respectively
Border level line, comprising:
Respectively by two corneal boundary image binaryzations, two corneal boundary bianry images are obtained；
The boundary for extracting two bianry images respectively, using the vertical tangent line in the outside on each boundary as corneal boundary reference line；
Corresponding corneal boundary image is adjusted separately according to two corneal boundary reference lines；
Two corneal boundary reference lines are determined again according to corneal boundary image adjusted.
10. the method according to claim 1, wherein obtaining corneal diameter according to two corneal boundary reference lines
Length, specifically:
The coordinate that two corneal boundary reference lines are respectively relative to respective corneal boundary image is converted in cornea image
Coordinate；
According to the lateral shift value of two corneal boundary reference lines coordinate in cornea image, the pixel span of corneal diameter is obtained
Value；
Corneal diameter length is obtained according to the pixel span value of corneal diameter, specifically:
If the physical length of corneal diameter is Lcornea, the pixel span value of corneal diameter is Pcornea, the green frame range of cornea image
Interior pixel quantity is Nimage, the inside casing area of scale material object is Aimage, then LcorneaFor
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711308312.9A CN109919896A (en) | 2017-12-11 | 2017-12-11 | A kind of measurement method of corneal diameter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711308312.9A CN109919896A (en) | 2017-12-11 | 2017-12-11 | A kind of measurement method of corneal diameter |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109919896A true CN109919896A (en) | 2019-06-21 |
Family
ID=66956778
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711308312.9A Pending CN109919896A (en) | 2017-12-11 | 2017-12-11 | A kind of measurement method of corneal diameter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109919896A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2497716Y (en) * | 2001-08-29 | 2002-07-03 | 江门市人民医院 | Instrument for measuring pupil |
| EP1754460A1 (en) * | 2005-08-19 | 2007-02-21 | Wavelight Laser Technologie AG | Presbyopia correction program |
| CN102988052A (en) * | 2011-09-09 | 2013-03-27 | 上海银晨智能识别科技有限公司 | Method and system for measuring foot length |
| CN103063143A (en) * | 2012-12-03 | 2013-04-24 | 苏州佳世达电通有限公司 | Measuring method and system based on image identification |
| CN202960457U (en) * | 2012-12-26 | 2013-06-05 | 牡丹江医学院 | Eyeglass frame type cornea diameter measurer |
| CN203474804U (en) * | 2013-09-25 | 2014-03-12 | 上海市嘉定区中心医院 | Object surface sampler |
| CN204618186U (en) * | 2015-04-08 | 2015-09-09 | 张青 | A kind of hand-held crossed cylinders |
| CN105205813A (en) * | 2015-09-06 | 2015-12-30 | 沈阳工业大学 | Cornea arcus senilis automatic detection method |
| CN107066957A (en) * | 2017-03-27 | 2017-08-18 | 北京理工大学 | Iris locating method and device in visible ray eyes image |
-
2017
- 2017-12-11 CN CN201711308312.9A patent/CN109919896A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2497716Y (en) * | 2001-08-29 | 2002-07-03 | 江门市人民医院 | Instrument for measuring pupil |
| EP1754460A1 (en) * | 2005-08-19 | 2007-02-21 | Wavelight Laser Technologie AG | Presbyopia correction program |
| CN102988052A (en) * | 2011-09-09 | 2013-03-27 | 上海银晨智能识别科技有限公司 | Method and system for measuring foot length |
| CN103063143A (en) * | 2012-12-03 | 2013-04-24 | 苏州佳世达电通有限公司 | Measuring method and system based on image identification |
| CN202960457U (en) * | 2012-12-26 | 2013-06-05 | 牡丹江医学院 | Eyeglass frame type cornea diameter measurer |
| CN203474804U (en) * | 2013-09-25 | 2014-03-12 | 上海市嘉定区中心医院 | Object surface sampler |
| CN204618186U (en) * | 2015-04-08 | 2015-09-09 | 张青 | A kind of hand-held crossed cylinders |
| CN105205813A (en) * | 2015-09-06 | 2015-12-30 | 沈阳工业大学 | Cornea arcus senilis automatic detection method |
| CN107066957A (en) * | 2017-03-27 | 2017-08-18 | 北京理工大学 | Iris locating method and device in visible ray eyes image |
Non-Patent Citations (3)
| Title |
|---|
| 余辉: "《Photoshop实用教程》", 30 June 2017, 重庆大学出版社, pages: 38 - 40 * |
| 洪汉玉: "《目标探测多谱图像复原方法与应用》", 31 October 2017, 国防工业出版社, pages: 155 * |
| 韩志艳: "《面向语音与面部表情信号的多模式情感识别技术研究》", 31 January 2017, 东北大学出版社, pages: 60 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9265414B2 (en) | Methods and systems for measuring interpupillary distance | |
| US10201273B2 (en) | Method for determining ocular measurements using a consumer sensor | |
| US9323075B2 (en) | System for the measurement of the interpupillary distance using a device equipped with a screen and a camera | |
| US11903644B2 (en) | Measuring eye refraction | |
| CN108846866B (en) | Optical imaging-based craniomaxillofacial soft tissue sagittal axial median plane determination method and system | |
| US20100091104A1 (en) | Systems and methods for the measurement of surfaces | |
| CN109008944A (en) | Sight measuring device, sight measuring program and sight measuring method | |
| US20150029322A1 (en) | Method and computations for calculating an optical axis vector of an imaged eye | |
| ES2754310T3 (en) | Aid procedure for determining the vision parameters of a subject | |
| CN107016697B (en) | A kind of height measurement method and device | |
| US20140240664A1 (en) | Method for measuring morpho-geometric parameters of a spectacle wearing individual | |
| US20150015848A1 (en) | Method for estimating a distance separating a pair of glasses and an eye of the wearer of the pair of glasses | |
| JP2012239566A (en) | Measuring apparatus for glasses, and three-dimensional measuring apparatus | |
| US20150049952A1 (en) | Systems and methods of measuring facial characteristics | |
| CN105354825A (en) | Intelligent device for automatically identifying position of reading material in read-write scene and application of intelligent device | |
| US20200394402A1 (en) | Object identification device, object identification system, object identification method, and program recording medium | |
| CN109900203A (en) | A kind of organism size non-contact measurement method | |
| CN105354822A (en) | Intelligent apparatus for automatically identifying position of read-write element in read-write scene and application | |
| CN112220444A (en) | Pupil distance measuring method and device based on depth camera | |
| WO2020072994A1 (en) | Calibration and image procession methods and systems for obtaining accurate pupillary distance measurement | |
| CN208740937U (en) | A kind of scale for non-contact measurement corneal diameter | |
| Wibirama et al. | Dual‐camera acquisition for accurate measurement of three‐dimensional eye movements | |
| CN109308472B (en) | Three-dimensional sight estimation method based on iris projection matching function | |
| CN105066901A (en) | Method for calculating eyeball shape, eye axis, and curvature of experiment animal by employing in-vitro measurement | |
| JP3711053B2 (en) | Line-of-sight measurement device and method, line-of-sight measurement program, and recording medium recording the program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |