JP2011250976A - Image processor, image processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress a resolution deterioration that occurs when correcting chromatic aberration of magnification including chromatic aberration of magnification generated in a fundus camera optical system and also that generated in an eyeball optical system which is an object of photography.SOLUTION: An imaging element 19 photographs a fundus image by using light reflected by the fundus and entering the imaging element via color filters of a plurality of colors. In accordance with focusing conditions of the fundus photography, an image signal processing section 23 converts, by using coordinates of a pixel value of a specific color among the plurality of colors as reference coordinates, coordinates of pixel values of other colors, and calculates, on the basis of pixel values of the other colors after coordinate conversion, pixel values of the other colors at coordinates before coordinate conversion which correspond respectively to the pixel values of the other colors after coordinate conversion. The image signal processing section 23 then interpolates, on the basis of the pixel value of the specific color, pixel values of the specific color at other coordinates, and interpolates, on the basis of the interpolated pixel values of the specific color and the pixel values of the other colors at the coordinates before coordinate conversion, pixel values of the other colors at the other coordinates.

Description

  The present invention relates to a technique for capturing a fundus image of an eye to be examined.

  A technique is known in which a correction function is created based on imaging conditions of an imaging optical system when medical image data is captured, and medical image data is corrected based on the correction function (see, for example, Patent Document 1). In such a conventional technique, correction of distortion occurring mainly in the fundus camera optical system has been the main focus. Here, since the distortion is a concentric deformation around the optical axis, the concentric circle is corrected with respect to the optical axis that is the center of the image. Further, in the case where astigmatism of the eye to be examined is considered as the imaging condition, Patent Document 1 describes that correction is axisymmetric with respect to the astigmatism axis, although this correction is not concentric (rotationally symmetric). Further, Patent Document 1 describes that it is possible to correct lateral chromatic aberration of a fundus camera by performing distortion on each color of RGB.

Japanese Patent Laid-Open No. 3-81879

  However, Patent Document 1 does not specifically describe a method for correcting the lateral chromatic aberration of the fundus camera. This correction is based on the assumption that all the pixels in the RGB image plane have been generated, and no correction method considering a pixel complementation process using a single-plate camera is mentioned. It was not always possible to maintain the resolution.

  Accordingly, an object of the present invention is to consider a pixel complementation process using a single-plate camera, and not only chromatic aberration of magnification that occurs in the fundus camera optical system but also chromatic aberration of magnification including that that occurs in the eyeball optical system that is the subject of photographing. This is to suppress a decrease in resolution that occurs when correction is made.

  An image processing apparatus according to the present invention includes an illuminating unit that illuminates the fundus of a subject's eye, an imaging unit that receives reflected light from the fundus through a plurality of color filters, and captures a fundus image, and imaging of the fundus A coordinate conversion means for converting the coordinates of pixel values of other colors using the coordinates of pixel values of a predetermined color of the plurality of colors as reference coordinates according to the in-focus state, and the other after the coordinate conversion Calculating means for calculating the pixel value of the other color at the coordinates before the coordinate conversion corresponding to the pixel value based on the pixel value of the color, and the pixel value of the predetermined color based on the pixel value of the predetermined color; First interpolation means for interpolating pixel values of other coordinates, pixel values of the predetermined color obtained by interpolating the pixel values by the first interpolation means, and the other in coordinates before coordinate conversion by the calculation means The other color based on the pixel value of the color of And having a second interpolation means for interpolating the pixel values of kicking other coordinates.

  According to the present invention, in consideration of the pixel complementation process using a single-plate camera, not only the lateral chromatic aberration that occurs in the fundus camera optical system, but also the lateral chromatic aberration that includes what occurs in the eyeball optical system that is the subject of photographing is corrected. Therefore, it is possible to suppress a decrease in resolution that occurs in the case of this.

It is a figure which shows the structure of the fundus camera which concerns on embodiment of this invention. It is a figure which shows the example of the display screen for performing a position and a focusing. It is a figure which shows the arrangement | positioning state of a general 3 color filter, and the spectral transmittance of each color filter of RGB. It is a figure for demonstrating the method of the pixel interpolation calculation in embodiment of this invention. It is a figure which shows the structure of an image signal process part. It is a figure for demonstrating the magnification chromatic aberration of a fundus image. It is a figure for demonstrating the gaze direction of the to-be-examined eye induced | guided | derived by the fixation light.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments to which the invention is applied will be described in detail with reference to the accompanying drawings.

  First, a first embodiment of the present invention will be described. FIG. 1 is a diagram illustrating a configuration of a fundus camera according to the first embodiment of the present invention. As shown in FIG. 1, the fundus camera according to the present embodiment has an aperture having a ring-shaped opening at the tip of the observation light source 1 on the optical path from the observation light source 1 and the imaging light source 3 to the objective lens 13. 5. A diaphragm 2 having a ring-shaped opening is disposed at the tip of the photographing light source 3 and reaches the mirror 4. A relay lens 6, a mirror 7, a relay lens 11, and a perforated mirror 12 are sequentially arranged at the tip of the mirror 4 to constitute a fundus illumination optical system 01. Note that the fundus camera according to the present embodiment has a configuration that is an application example of the image processing apparatus of the present invention.

  In the reflection direction of the mirror 7, a two-hole aperture 36, a lens 8, a focusing index 9, and a focusing index light source 10 are arranged to constitute a focusing index projection optical system 03. The focusing index projection optical system 03 moves in the direction A in the figure in conjunction with the focusing lens 14, projects the focusing index on the fundus Er of the eye to be examined, and in the direction B in the figure when taking a still image. Move and retract from the illumination optical system 01.

  On the optical path in the transmission direction of the perforated mirror 12, a focusing lens 14, a photographing lens 15, and an imaging unit 20 are disposed. An oblique partial reflection mirror 38 is disposed between the photographic lens 15 and the imaging unit 20. In order to fix the line of sight of the subject's eye, a fixation lamp optical path for fixing the subject's eye is prepared. A plurality of fixation lamp light sources 39 formed of LED arrays are arranged at positions substantially conjugate with the imaging section 20 of the fixation lamp optical path, and are connected to the control section 25. The fixation lamp light source 39 is controlled so that one LED at the position input by the fixation lamp position change switch 36 is lit. The focusing lens 14 is connected to a lens position detector 37 that can detect the lens position, and the output of the lens position detector 37 is also connected to the control unit 25. An alignment index light source 18 is connected to the perforated mirror 12 through an optical fiber 17. The above configuration is the fundus photographing optical system 02.

  An imaging element 19 and a three-color wavelength decomposing unit 16 are disposed in the imaging unit 20, and an actual mask 40 that defines an imaging range is bonded immediately before the three-color wavelength decomposing unit 16. Further, the output of the imaging unit 20 is sequentially connected to an image signal processing unit 23 and a display 24 for displaying an image. The observation light source 1 is connected to the observation light source drive circuit 27. The photographing light source 3 is connected to the photographing light source driving circuit 26. The focus index light source 10 is connected to the focus index light source drive circuit 22. The alignment index light source 18 is connected to the alignment index light source driving circuit 21. Note that the alignment index light source drive circuit 21, the focusing index light source drive circuit 22, the image signal processing unit 23, the imaging light source drive circuit 26, the observation light source drive circuit 27, the input unit 28, the recording unit 34, and fixations. The lamp position change switch 36 is connected to the control unit 25.

  Next, operations during fundus observation will be described. The control unit 25 drives the observation light source drive circuit 27 to turn on and dim the observation light source 1 and lights one of the fixation lamp light sources 39 as a fixation lamp. The imaging region on the fundus can be determined by having the subject gaze at the fixation lamp. The light beam emitted from the observation light source 1 passes through a diaphragm 5 and a mirror 4 having a ring-shaped opening. Here, the observation light source 1 is composed of a near-infrared LED having a center wavelength at 850 nm, and the mirror 4 is a dichroic mirror that transmits infrared light and reflects visible light. The emitted light beam passes through the mirror 4. The light beam that has passed through the mirror 4 passes through the relay lens 6, the mirror 7, and the relay lens 11, is reflected around the perforated mirror 12, and illuminates the fundus Er through the objective lens 13, the cornea Ec of the eye E to be examined, and the pupil Ep. To do.

  The control unit 25 drives the focusing index light source driving circuit 22 to turn on the focusing index light source 10. The focusing index light source 10 illuminates the focusing index 9, and the image of the focusing index 9 passes through the lens 8, is reflected by the mirror 7, and is superimposed on the light beam from the observation light source 1 to be examined E Projected onto the fundus Er. The focusing index light source 10 is composed of a near-infrared LED having a center wavelength at 850 nm.

  The illuminated fundus image and focusing index image pass through the pupil Ep, the cornea Ec, the objective lens 13, and the apertured mirror 12 of the eye E, and pass through the focusing lens 14 and the imaging lens 15. The light passes through the three-color wavelength resolving unit 16 in the imaging unit 20 and is incident on the imaging device 19 to form an image.

  The controller 25 drives the alignment index light source driving circuit 21 to turn on the alignment index light source 18. The alignment index light source 18 irradiates the cornea Ec of the eye E through the optical fiber 17 and the objective lens 13. The reflected light is superimposed on the reflected light from the fundus of the observation light source 1 and the focusing index light source 10 and is imaged on the image sensor 19. Here, the alignment index light source 18 is composed of a near-infrared LED having a center wavelength of 850 nm.

  In the image sensor 19, photoelectric conversion is performed on the formed fundus image, the focusing index image, and the alignment index image. The image signal processing unit 23 reads and amplifies data from the image sensor 19. Thereby, digital image data which is a moving image is generated and displayed on the display unit 24. This is shown in FIG. FIG. 2A shows a state in which the position and focus are shifted. The operator moves the fundus camera back and forth, right and left, and up and down by an operation unit (not shown) while viewing this image, and the fundus of the eye to be examined. Align the image. Further, the focusing index projection optical system 03 is moved in the direction of A in the figure in conjunction with the focusing lens 14 to perform focusing. FIG. 2B shows a state in which the alignment index is aligned with the alignment index circle, the alignment is completed, and the focusing index is in a straight line and the focusing is completed. A hatched portion 40 ′ outside the fundus image is an image of the substance mask 40 that defines the imaging range adhered immediately before the three-color wavelength separation unit 16. It has the role of cutting only the fundus image in the range where flare and ghost are not mixed by cutting off the reflected light and scattered light generated in the anterior segment of the eye to be examined.

  Here, the configuration of the imaging unit 20 and the operation during observation will be described in detail. The imaging device 19 in the imaging unit 20 has red (hereinafter referred to as “R”), green (hereinafter referred to as “G”), blue (hereinafter referred to as “R”) mosaic arrangement so as to match each pixel of the imaging device 19. , B)) is arranged. The image signal processing unit 23 reads and amplifies the data of each pixel from the image sensor 19, calculates all pixel values, and generates image data.

  Here, the calculation of each pixel value at the time of image data generation will be described in detail. FIG. 3A shows an arrangement state of a general three-color color filter. In each pixel of the imaging unit 20, RGB color filters having the spectral transmittance shown in FIG. 3B are arranged, and the corresponding pixels are assigned to the color filters. Each filter prepared for each pixel is sensitive to near infrared light (here, 850 nm) having a wavelength longer than 700 nm, which is the visible range, and the fundus observation described above is performed using this output. become. In general, a pixel value (G) of a specific color element of a pixel assigned to another color element (for example, RB) is created from pixels allocated to the specific color element (for example, G), and each color of RGB Plane image data is generated. On the other hand, in this embodiment, the wavelength used at the time of observation is 850 nm, and each RGB color filter has substantially the same transmittance at that wavelength. Therefore, the image signal processing unit 23 may sequentially read out each pixel and produce a monochrome image signal as it is.

  Next, the operation at the time of fundus photographing will be described. The operator aligns and focuses while viewing an image as shown in FIG. 2A displayed on the display 24, and when the position and the focus are matched, a shooting switch configured in the input unit 28 is set. Push. As a result, the control unit 25 causes the lens position detector 37 to detect the position of the focus lens 14 and drives the alignment index light source drive circuit 21 and the focus index light source drive circuit 22 to focus the index light source for focus. 10. The alignment index light source 18 is turned off. Then, the control unit 25 drives the focusing index projection optical system 03 in the direction B and retracts it outside the optical path, and then drives the photographing light source drive circuit 26 to cause the photographing light source 3 to emit light.

  The light beam emitted from the imaging light source 3 passes through the diaphragm 2 having a ring-shaped opening, is reflected by the mirror 4, and thereafter illuminates the fundus Er of the eye E to be examined along the same path as the observation light source. Then, the fundus image, which is the reflected light, is guided to the image sensor 19 to form an image. Here, since the imaging light source 3 is visible light and the mirror 4 is a dichroic mirror that transmits infrared light and reflects visible light, only the visible light region emitted from the imaging light source 3 is reflected by the mirror 4. reflect.

  The image sensor 19 performs photoelectric conversion, and the image signal processing unit 23 generates fundus image data composed of RGB plane image data as a still image, which is displayed on the display unit 24 and simultaneously recorded via the control unit 25. Fundus image data is recorded in the unit 34. The concept of the operation of the image signal processing unit 23 at that time will be described in detail with reference to FIGS.

  FIG. 5 is a diagram showing a detailed configuration of the image signal processing unit 23. The pixel data read by the reading unit 231 from the image sensor 19 having the pixel arrangement shown in FIG. 3A passes through the calculation unit 232, and the R plane memory 235 corresponding to each RGB color provided in the virtual memory 234. The G plane memory 236 and the B plane memory 237 are arranged. The state of the pixel values on each plane memory at this time is shown in 401 to 403 in FIG. 401 indicates the arrangement of the R plane memory 235, 402 indicates the arrangement of the G plane memory 236, and 403 indicates the arrangement of the B plane memory 237. That is, 401 to 403 in FIG. 4 show a state in which the pixel value of the pixel is embedded at the pixel position where the filter corresponding to each color element is arranged. At this time, pixel values (for example, R pixel value at the coordinate position of G01 on the R plane) for pixels captured with other color elements are still blank.

  Next, the computing unit 232 stores a coordinate conversion function corresponding to the diopter of the eye to be examined corresponding to the focus lens position detected at the time of photographing from the correction storage unit 233 storing the coordinate correction position for correcting the chromatic aberration of magnification. In other words, the coordinate conversion function of the other color element (corresponding to RB in this embodiment) with respect to the reference color element (corresponding to G having the largest number of pixels in this embodiment) is read and arranged. The corrected coordinates are calculated. Further, the calculation unit 232 calculates a pixel value corresponding to the original reference coordinate from surrounding pixel values having the corrected coordinates, and rewrites the pixel value. This concept is indicated by 404 and 406 in FIG. That is, the determination of the pixel value of R22 (denoted as R22 ′ in 404 of FIG. 4) at the time of 401 in FIG. 4 is determined on the same meridian with respect to the origin of the coordinate axes, that is, O which is the center of the imaging screen. And R44 are once subjected to coordinate conversion located at R22 and R44 indicated by 404 in FIG. 4 to obtain corrected coordinates of R22 and R44. At this time, the coordinates of R00 are the corrected reference coordinate values, the coordinates of R00 ′ are the original reference coordinate values, and the distance ΔB between them corresponds to the amount of lateral chromatic aberration to be corrected. Thereafter, the calculation unit 232 calculates the pixel value R22 ′ on the original reference coordinates by linear interpolation calculation or the like from the coordinates after the coordinate conversion of both. In the description, for simplicity, R22 and R44 on the same meridian have been described as examples. However, if they are not on the same meridian, they are calculated from four peripheral pixels surrounding the symmetric pixel in the vicinity of the symmetric pixel. It will be good. The concept of pixel interpolation for the B plane is the same, but in 406 in FIG. 4, the R plane indicated by 404 in FIG. 4 is imaged at a distance close to the origin due to lateral chromatic aberration, but shifted in the reverse direction. It shows that.

  Image data in which the lateral chromatic aberration is corrected by the above calculation is arranged on each image plane. The coordinate conversion function prepared in the correction storage unit 253 includes a plurality of functions created as follows, for example.

First, a method for creating a reference (straight vision) coordinate conversion function will be described.
An eyeball optical system model having standard chromatic aberration and design data of a fundus camera optical system arranged so that its optical axis coincides with it are prepared. On the other hand, the calculation unit 232 is based on multiple ray tracing composed of a plurality of wavelengths having a spectral distribution weight that takes into account the standard spectral distribution of the light source used, the transmittance of the optical system, and the color filter characteristics of the three-color separation means. The center of gravity position of each point image of RGB at each image position is obtained by point image intensity distribution simulation. The difference between the obtained RB point image centroid position and the reference G point image centroid position is the correction amount. In the present embodiment, the center of gravity of the point image intensity distribution is used for the calculation of the correction value. However, the median value may be adopted or may be used in combination with RB. Further, the correction value can be determined so that the maximum value of the combined point image intensity distribution is maximized. At this time, it is sufficient to calculate a plurality of image heights and determine by interpolation between them without having to calculate all the continuous image heights. In addition, this interpolation may be calculated in advance and prepared in the correction storage unit 253 in the LOT format corresponding to each pixel, or a method in which the sequential calculation unit 232 calculates when actually correcting. Furthermore, although the difference itself from the point image center of gravity of G may be used as a correction amount, seasoning such as a small estimate in advance is possible in order to avoid overcorrection due to individual differences in the eye to be examined. Naturally, since there are individual differences in the eye to be examined, a strict correction amount is not necessary, and even if such a seasoning is performed, although there is a difference in degree, it does not change that it has an image improvement effect . What is important is not to target chromatic aberration that occurs only in the optical system of the fundus camera, but to make corrections that also take into account the optical system of the subject's eye that is the subject.

Next, a method for creating a coordinate transformation function for the eye diopter (in-focus state) will be described.
An eyeball optical system model when the subject's eye is not normal is not established. Therefore, in this embodiment, the amount of correction of the necessary chromatic aberration of magnification is estimated from human eye images of various eye diopters that are actually captured by a fundus camera having a known optical design value and an imaging means having a known filter design value. We use what we did. FIG. 6 shows an example of a method for determining the correction value. FIG. 6A is a color image that is actually taken with the optical axis substantially aligned with the eye to be examined whose diopter is known. Here, FIG. 6B is a plot of the pixel values of each RGB plane in the region of the arrow A on the meridian of the blood vessel perpendicular to the meridian from the center O of the screen, and the blood vessel is shown here because it is dark red. The profile is as follows. The positions of the blood vessels are the bottoms Bb, Gb, and Rb of each plot. If there is no lateral chromatic aberration, this should be almost the same. Accordingly, the aberration amounts to be corrected on the R plane and the B plane are ΔB and ΔR, respectively. This can also be measured in FIG. 6B at different angles of view to obtain chromatic aberration at each angle of view. By performing such image measurement on a plurality of eyes to be examined having substantially the same diopter and obtaining an average thereof, the correction amount at each diopter can be determined. At this time, it is sufficient to calculate for a plurality of diopters and determine by interpolation between them without having to calculate all the continuous diopters. In addition, this interpolation may be calculated in advance and prepared in the correction storage unit 253 in the LOT format corresponding to each pixel, or a method in which the sequential calculation unit 232 calculates when actually correcting may be used. This is the same as the method for determining the correction amount for the corner. Note that the above-described reference (straight vision) coordinate conversion function may also be calculated from a fundus image that is actually captured.

  The above coordinate conversion is only for the fundus image located inside the actual mask image 40 ′ defining the imaging range adhered immediately before the three-color wavelength separation unit 16 shown in FIG. 2. The chromatic aberration of magnification occurs only in the fundus image, and since this substantial mask is imaged without any eyeball optical system or imaging optical system, if the same correction as the fundus image is performed, the mask This is because color blur as an artifact is generated at the boundary of the image and becomes diagnostic noise. Of course, the calculation unit 232 does not use the entity mask 40 for simplification of the calculation, performs coordinate conversion on the entire imaging range, and performs pixel interpolation processing on the RGB plane described later, and corresponds to the entity mask. The same effect can be obtained even if the process of painting the pixels of the mask portion to be blacked out is performed.

Next, the calculation unit 232 performs G-plane pixel interpolation on the pixel values of the G-plane memory 236, and stores the pixel values of blank pixels in the G-plane memory 236. A method of calculating a blank pixel value on each color plane coordinate performed by the calculation unit 232 by pixel interpolation will be described by taking a pixel surrounded by a black frame 407 to 409 in FIG. 4 as an example. As indicated by reference numeral 408 in FIG. 4, the blank pixel of G22 is surrounded by pixels that already have pixel values on the top, bottom, left, and right. Therefore, interpolation is possible by taking the average of the four pixels as follows.
G22 = (G12 + G21 + G23 + G32) / 4
The same calculation is possible for G11, G13, and the like.

Next, the calculation unit 232 performs pixel interpolation on the R plane. There are three types of R blank pixels. That is, R12 and R32 are sandwiched between pixels that already have pixel values. R21 and R23 are sandwiched between pixels that already have pixel values. Further, R11 is surrounded by four pixels having already pixel values in the oblique direction. It is naturally possible to perform pixel interpolation using these, but the amount of information to be used is small compared to interpolation using pixels close to the top, bottom, left, and right as in G. Therefore, this interpolation increases the possibility of losing high-frequency components in the image space, and the resolution is not maintained. Therefore, in this embodiment, interpolation using a color difference signal using pixel values on the G plane is performed to maintain this high frequency component. That is, blank pixel values are determined by interpolation calculation for the three types of R pixels as follows, and writing to the R plane memory is performed.
R12 = ((R02′−G02) + (R22′−G22) −2xG12) / 2
R21 = ((R20′−G20) + (R22′−G22) −2xG21) / 2
R11 = ((R00'-G00) + (R02'-G02) + (R20'-G20) + (R22'-G22) -4xG11) / 4
The B plane interpolation is performed in the same manner as described below, and the result is stored in the B plane memory.
B12 = ((B11′−G11) + (B13′−G13) −2xG12) / 2
B21 = ((B20−G20) + (B22−G22) −2xG21) / 2
B22 = ((B11′−G11) + (B13′−G13) + (B31′−G31) + (B33′−G33) −4xG22) / 4

  As described above, in the present embodiment, the calculation unit 232 maintains the image quality by using the high resolution of the plane with the largest number of pixels (G plane in the present embodiment) in the second interpolation calculation. I am trying. In other words, information on two color element planes, that is, the color element plane and the reference color element plane, is used for interpolation of the color element plane that is not the reference color element plane.

  From the above description, it will be understood that the second interpolation calculation for the R plane and the B plane should not be performed before the first interpolation calculation for the G plane performed after the coordinate conversion described above. Even in a method other than this embodiment, when performing pixel interpolation from a plurality of color element planes instead of a single color element plane, it goes without saying that the order of the two interpolations is important. . Comparing with a general virtual pixel calculation method, that is, an interpolation calculation using three 4 × 4 color elements, the merits of this method become clear.

In a general interpolation calculation, the values of RGB colors of the virtual pixel P00 in 407 to 409 in FIG. , R22, B31, and B33 are calculated from the received light data by the following formula. If the RGB values of the virtual pixels are R _P00 , G _P00 and B _P00 , respectively, the following is obtained.
G _P00 = (G12 + G21) / 2
R _P00 = (9 · R22 + 3 · R02 + 3 · R20 + R00) / 16
B _P00 = (9 · B11 + 3 · B13 + 3 · B31 + B33) / 16

  It will be understood that although the pixels used during pixel interpolation are weighted, averaging is performed using pixels at a considerable distance. After the calculation unit 232 finishes all calculations, the reading unit 238 reads all image data from the virtual memory 234 and transfers the image data to the control unit 25. The controller 25 uses this for display and recording.

  In the description of the present embodiment, the calculation performed for each color element plane has been described. However, the calculation performed by the actual calculation unit is not necessarily limited to this order, and an algorithm that performs calculation repeatedly for each small area is adopted. In addition, the calculation time can be shortened. Further, when performing display immediately after shooting, the process of correcting the chromatic aberration of magnification can be omitted to enable display in a short time after shooting. In such a case, it is desirable to execute this calculation by background processing and switch the display after the calculation is completed.

  According to the present embodiment, in consideration of the pixel complementation process using a single-plate camera, not only the lateral chromatic aberration that occurs in the fundus camera optical system, but also the lateral chromatic aberration that includes what occurs in the eyeball optical system that is the object to be imaged. It is possible to suppress a decrease in resolution that occurs when correction is performed.

  Next, a second embodiment of the present invention will be described. In the first embodiment, an eyeball optical system model having standard chromatic aberration and design data of a fundus camera optical system arranged so that its optical axis coincides are prepared as a reference coordinate conversion function for correcting lateral chromatic aberration. . However, in recent years, for cataractous eyes, an operation in which an opaque lens is excised and an artificial lens IOL is inserted has become widespread. It is known that the lateral chromatic aberration of an eye to be examined (so-called IOL eye) on which such a surgery has been performed is different from the standard lateral chromatic aberration. The configuration of the fundus camera according to the present embodiment is the same as that of the fundus camera according to the first embodiment shown in FIG.

  Therefore, in the second embodiment of the present invention, such a coordinate conversion function for the IOL eye and an IOL eye selection switch for inputting that the examiner determines that the eye to be examined is an IOL eye are provided. ing. The IOL eye selection switch is connected to the control unit 25 in FIG. Further, when the control unit 25 detects that the IOL eye selection switch is turned on, the calculation unit 232 in FIG. 5 performs another operation corresponding to the IOL eye from the correction storage unit 253 storing the coordinate conversion function for correcting the chromatic aberration of magnification. After calling the coordinate conversion function, the same calculation as in the first embodiment is performed. When the eye to be examined is not normal, the coordinate conversion function corresponding to the in-focus lens position detected at the time of shooting is the same as in the first embodiment, and other configurations and operations are the same as in the first embodiment. Therefore, explanation is omitted.

  The reference coordinate conversion function for the IOL eye prepared here is created in the same manner as the reference coordinate conversion function for the normal eye to be examined by using an eyeball optical system model having standard chromatic aberration and replacing the lens portion with IOL. Also, the method for creating the coordinate conversion function for each diopter is the same except that the IOL eye is used as the eye to be examined. However, although the reference coordinate conversion function is different, the change of the coordinate conversion function with respect to the diopter (change in lateral chromatic aberration) itself is substantially the same as the change for the standard eye to be examined. Therefore, instead of preparing the reference coordinate conversion function for the IOL eye, a correction coordinate conversion function for the IOL eye may be prepared to perform the actual coordinate conversion calculation. In other words, the IOL eye may be first subjected to the coordinate conversion corresponding to the diopter of the standard eye and then the correction coordinate conversion for the IOL eye may be performed. According to this method, the number of coordinate conversion functions or coordinate conversion LOTs stored in the correction storage unit 253 can be reduced.

  Next, a third embodiment of the present invention will be described. In the first and second embodiments, as shown in FIG. 7A, an example has been described in which correction is made point-symmetrically with respect to the optical axis such that the optical axis of the eye to be examined coincides with the imaging optical axis. However, in actual photographing, as described in the first embodiment, the examiner operates the fixation lamp position change switch 36 shown in FIG. As a result, one of the fixation lamp light sources 39 is turned on as a fixation lamp, and, as shown in FIG. decide. At this time, the line of sight (inspection eye optical axis) SL of the subject guided by the fixation lamp has an angle of θ in the meridian direction and φ in the tangential direction with respect to the optical axis of the objective lens. In such a case, the optical axis of the eye to be inspected does not coincide with the optical axis of the objective lens 13, and an asymmetric magnification chromatic aberration is generated in the photographing optical axis due to the inclination of the eye optical system to be examined. The function is a function of the line-of-sight direction θ of the eye, and the direction is a function of the line-of-sight direction φ. In FIG. 7, E is the eye to be examined, Er is the fundus imaging region of the eye to be examined, and Ep is the eye to be examined. The configuration of the fundus camera according to the present embodiment is the same as that of the fundus camera according to the first embodiment shown in FIG. In addition, the fundus camera according to the present embodiment detects the subject's line of sight (the optical axis of the eye to be examined) from the position of the fixation lamp.

  Therefore, in this embodiment, a plurality of coordinate conversion functions for correcting such asymmetric chromatic aberration of magnification with respect to the optical axis are prepared in the correction storage unit 253. When the examiner operates the fixation lamp position change switch 36, the control unit 25 controls the light emission position of the fixation lamp, and calculates a coordinate conversion function for correcting asymmetric magnification chromatic aberration corresponding to the fixation lamp position. The operation unit 232 is instructed to be called from the correction storage unit 253. Then, the calculation unit 232 calls the coordinate conversion function for correcting the chromatic aberration of magnification that is asymmetric with respect to the optical axis from the correction storage unit 253, and then performs the same calculation as in the first embodiment. When the eye to be examined is not normal, the coordinate conversion function corresponding to the in-focus lens position detected at the time of shooting is the same as in the first embodiment, and other configurations and operations are the same as in the first embodiment. Therefore, explanation is omitted.

  The coordinate correction function for correcting the asymmetric magnification chromatic aberration prepared here is an arrangement in which the eyeball optical system model having standard chromatic aberration is tilted in the direction of the line of sight θ guided by the fixation lamp with respect to the optical axis of the photographing optical system. It can be created by simulating in the same manner as the reference coordinate conversion function for the normal eye to be examined. Since the eyeball optical system and the photographing optical system are basically composed of coaxial optical elements, the coordinate transformation function may be rotated in the line-of-sight direction φ with respect to the φ direction. Also, a method for creating a coordinate conversion function for each diopter can be created from a photographed image as in the first embodiment.

  However, the change in the chromatic aberration of magnification with respect to the change in the line-of-sight direction has a characteristic that occurs in addition to the reference coordinate conversion function in the first embodiment. Therefore, when the line of sight is changed, first, coordinate conversion corresponding to the diopter of the standard eye is performed, and then the coordinate conversion of the predetermined diopter with respect to the line of sight θ is performed in the direction of the line of sight φ. You can take steps to apply. According to this method, the number of coordinate conversion functions or coordinate conversion LOTs stored in the correction storage unit 253 can be reduced.

  Furthermore, in the present embodiment, the position of the fixation lamp is used in the detection of the gaze method, but it goes without saying that the gaze direction can also be detected by discriminating between the nipple position on the actual image and the left and right eyes. Absent. In this case, if the presence of the nipple cannot be confirmed on the image, the lateral chromatic aberration can be corrected.

  According to the above embodiment, it is possible to correct lateral chromatic aberration while maintaining a sufficient resolution of the image sensor. Since the fundus image generated by the above embodiment has almost no chromatic aberration and has a sufficient resolution, it has a very high diagnostic value as compared with the conventional fundus image. Further, by setting the reference plane to the G plane in the pixel interpolation process, it is possible to share the same signal processing with the conventional single-panel color camera, and the cost can be reduced.

  The coordinate conversion function used here is generated using design data of the fundus camera optical system and a correction amount of lateral chromatic aberration generated by an eyeball optical system model having standard chromatic aberration. Alternatively, the coordinate conversion function is generated by using a correction amount of lateral chromatic aberration estimated from images of various diopter diopters photographed with a known optical system. Accordingly, it is possible to sufficiently correct the lateral chromatic aberration including the lateral chromatic aberration that occurs in the optical system of the eye to be examined.

  Furthermore, in the above-described embodiment, by providing a coordinate conversion function according to the output of the lens position detector 37, it is possible to perform appropriate correction even for the subject's eyes having different diopters. In the above-described embodiment, a fundus image free from artifacts generated in the mask portion can be provided by applying an electronic mask to the generated fundus image. In addition, the same effect can be obtained by limiting the coordinate transformation to the passage part of the substance mask.

  Moreover, in the said embodiment, if it inputs that it is an IOL eye, correction | amendment with respect to an IOL eye can also be performed appropriately. Furthermore, in the above-described embodiment, appropriate correction can be performed for any imaging region by giving a coordinate conversion function corresponding to the line-of-sight direction (optical axis) of the eye to be examined. By using the fixation target position for detection of the gaze direction at this time, this effect can be obtained without using a special configuration. The line-of-sight direction can also be estimated by discriminating the nipple position and the left and right eyes.

  The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed.

  23: Image signal processing unit, 25: Control unit, 37: Lens position detector, 39: Fixation lamp light source, 232: Calculation unit, 235: R plane memory, 236: G plane memory, 237: B plane memory

Claims (12)

Illumination means for illuminating the fundus of the eye to be examined;
An imaging unit that receives reflected light from the fundus via a color filter of a plurality of colors and captures a fundus image;
Coordinate conversion means for converting the coordinates of pixel values of other colors using the coordinates of pixel values of a predetermined color of the plurality of colors as reference coordinates according to the in-focus state at the time of photographing the fundus;
Based on the pixel value of the other color after coordinate conversion, calculation means for calculating the pixel value of the other color at coordinates before coordinate conversion corresponding to the pixel value;
First interpolation means for interpolating pixel values of other coordinates in the predetermined color based on the pixel values of the predetermined color;
Based on the pixel value of the predetermined color in which the pixel value is interpolated by the first interpolation means and the pixel value of the other color in the coordinates before coordinate conversion by the calculation means, the other color in the other color And a second interpolation means for interpolating pixel values of the coordinates of the image processing apparatus.   The image processing apparatus according to claim 1, wherein the predetermined color is a color having the largest number of pixels among the plurality of colors in the imaging unit.   The coordinate conversion means includes a coordinate conversion function generated using an eyeball optical system model having standard chromatic aberration and design data of a fundus camera optical system arranged so that an optical axis coincides with the eyeball optical system model. The image processing apparatus according to claim 1, wherein coordinate conversion is performed using the image processing apparatus.   The coordinate conversion means performs coordinate conversion using a coordinate conversion function generated based on a correction amount of chromatic aberration of magnification estimated from a human eye image of a plurality of eye diopters photographed with a known optical system. The image processing apparatus according to claim 1, wherein: Mask means for applying a mask for defining an imaging range of the imaging means;
5. The image processing apparatus according to claim 1, wherein the coordinate conversion unit performs coordinate conversion on a pixel value other than a portion on which the mask is applied. 6.   The image processing apparatus according to claim 1, wherein the coordinate conversion unit performs coordinate conversion corresponding to an IOL eye.   6. The image processing according to claim 1, wherein the coordinate conversion unit performs coordinate conversion corresponding to a case where the optical axis of the eye to be examined does not coincide with the optical axis of the imaging unit. apparatus.   The image processing apparatus according to claim 7, wherein the coordinate conversion unit performs different coordinate conversion according to a direction of an optical axis of the eye to be examined.   The image processing apparatus according to claim 8, further comprising detection means for detecting a direction of an optical axis of the eye to be examined from a position of a fixation lamp.   The image processing apparatus according to claim 8, further comprising a detecting unit that detects a direction of an optical axis of the eye to be examined from a nipple position and a left / right eye distinction. An image processing method executed by an image processing apparatus including an illuminating unit that illuminates the fundus of a subject's eye and an imaging unit that receives reflected light from the fundus via a plurality of color filters and captures a fundus image There,
A coordinate conversion step for converting the coordinates of pixel values of other colors using the coordinates of pixel values of a predetermined color of the plurality of colors as reference coordinates according to the in-focus state at the time of photographing the fundus;
Based on the pixel value of the other color after coordinate conversion, a calculation step of calculating the pixel value of the other color at coordinates before coordinate conversion corresponding to the pixel value;
A first interpolation step of interpolating pixel values of other coordinates in the predetermined color based on the pixel values of the predetermined color;
Based on the pixel value of the predetermined color in which the pixel value is interpolated by the first interpolation step and the pixel value of the other color at the coordinates before coordinate conversion by the calculation step, the other color in the other color And a second interpolation step for interpolating pixel values of the coordinates of the image processing method. An image processing method executed by an image processing apparatus comprising: an illuminating unit that illuminates the fundus of the eye to be examined; and an imaging unit that receives reflected light from the fundus via a plurality of color filters and captures a fundus image. A program for causing a computer to execute,
A coordinate conversion step for converting the coordinates of pixel values of other colors using the coordinates of pixel values of a predetermined color of the plurality of colors as reference coordinates according to the in-focus state at the time of photographing the fundus;
Based on the pixel value of the other color after coordinate conversion, a calculation step of calculating the pixel value of the other color at coordinates before coordinate conversion corresponding to the pixel value;
A first interpolation step of interpolating pixel values of other coordinates in the predetermined color based on the pixel values of the predetermined color;
Based on the pixel value of the predetermined color in which the pixel value is interpolated by the first interpolation step and the pixel value of the other color at the coordinates before coordinate conversion by the calculation step, the other color in the other color A program for causing a computer to execute a second interpolation step for interpolating pixel values of the coordinates.

Images